On September 30, Alphabet's Google announced the launch of a highly advanced artificial intelligence model named "Argon" as part of its Gemini 4 lineup, signaling a concerted effort to regain ground in the competitive AI landscape alongside players like Anthropic and OpenAI.

Company representatives described Argon as being notably larger than its predecessor, the "Pro" models, and claimed it is the most capable model to date, specially designed for handling intricate tasks. They further asserted that its performance can be compared to leading models such as OpenAI's Astra and Anthropic's Opus, particularly in key coding and cybersecurity metrics.

Google intends to share Argon with a select group of cybersecurity partners and is engaged in a voluntary initiative from the Trump administration that allows pre-release access to certain models. However, the company did not specify when the model might be made available to the general public.

According to Google's press materials, Argon delivered superior results compared to Astra and Opus on various standardized performance tests. Yet, there remain areas where it did not meet expectations, particularly in two out of four coding benchmarks highlighted in the release.

In a significant shift in its product strategy, Google has decided against releasing the previously anticipated Gemini 3.5 Pro, which CEO Sundar Pichai had indicated would debut in June.

Meanwhile, Google has undergone a restructuring of its DeepMind AI division, with founder Demis Hassabis stepping down and several leaders associated with the Gemini model's development departing from the organization.

These recent delays in the rollout of AI models have left Google trailing behind competitors like Anthropic and OpenAI, which have continually updated and enhanced their top-tier AI offerings. In response to concerns about losing its competitive edge, Pichai stated in July that Google was not falling behind in the AI race, though the company's messaging has shifted towards emphasizing not just cutting-edge technology, but also the cost-effectiveness of its models compared to those of its rivals.

Google unveils its flagship AI model Gemini 4 following extended delays.

A sunlit Google research lab with floor-to-ceiling windows bathing the room in warm morning light; at its center, a sleek metallic pedestal holds a polished “Gemini 4” module, its name etched in fine lettering. A small group of professional engineers in smart-casual attire admire it as the Google logo hangs prominently on the wall behind them.

Google announces Gemini 4 flagship AI model after months of delays - Reuters

- Google announced the new "Argon" AI model to compete with OpenAI and Anthropic.  
- Argon is larger and performs well on key coding and cyber industry benchmarks.  
- Google will not release Gemini 3.5 Pro amid ongoing delays and staffing changes.  

SAN FRANCISCO, Sept 30 (Reuters) - Alphabet's Google (GOOGL.O) , opens new tab on Wednesday announced a new top-tier artificial intelligence model to ​anchor its Gemini 4 generation of models in a fresh bid ‌to catch up to rivals Anthropic and OpenAI in the AI race.

The new model, "Argon", is larger in size than Google's previous line of advanced "Pro" models, a company spokesperson ​said.

Sign up here.

"It's our most performant model yet built for complex workloads, ​and we see it comparable to frontier models like (OpenAI's) Astra and (Anthropic's) Opus on ⁠key coding and cyber benchmarks," the spokesperson said.

Google said it was ​providing the model to select cybersecurity partners and was participating in the Trump ​administration's voluntary process for pre-release model access . It did not provide a timeline for the model's public release.

The model posted better marks than Astra and Opus in terms of self-reported ​performance on several industry benchmarks, per a company press release. Still, Argon ​remained behind on certain other metrics, including two of the four coding-related benchmarks Google ‌included ⁠in the release.

Google no longer plans to release Gemini 3.5 Pro, which CEO Sundar Pichai originally said would be released in June, according to the company spokesperson.

In the intervening time, Google overhauled its DeepMind AI lab, with its ​founder and CEO ​Demis Hassabis stepping ⁠aside and several leaders of its Gemini model leaving the company.

The months of delays to its AI model have ​caused the company to fall behind Anthropic and OpenAI, ​which continued ⁠to push the research frontier with new releases to its top models.

Pichai in July pushed back on the notion that Google was losing ground in the ⁠AI ​race. But Google in recent months has shifted ​its messaging from touting Gemini's cutting-edge capabilities to highlighting its cost advantage compared to competitor models.

Adds details and context in paragraphs 5-9

Reporting ​by Kenrick Cai in San Francisco; Editing by Nick Zieminski and Sonali Paul

Our Standards: The Thomson Reuters Trust Principles. , opens new tab

X

Facebook

Linkedin

Email

Link Purchase Licensing Rights

Google announces Gemini 4 flagship AI model after months of delays

Assessing the significance of an advertisement can be crucial for viewers. Several factors can affect the experience, including the video player’s performance. For instance, if the video player is sluggish, it can hinder the loading of content, leading to frustration. Additionally, instances where the video content fails to load or freezes can disrupt the viewing experience entirely. Other issues may arise, such as the video not commencing after the advertisement finishes or the audio being excessively loud, further impacting how relevant and engaging the ad feels. Addressing these concerns is essential for advertisers aiming to connect effectively with their audience.

Is China primarily utilizing AI tokens for creating micro-dramas?

A rooftop in a modern Chinese city at sunrise, warm natural light bathing a cascade of golden “AI” tokens falling onto a row of smartphones, each screen showing a tiny intense drama scene; a single Chinese flag waves gently in the background, cinematic composition.

Is China using AI tokens mostly to produce micro-dramas? - CNN

- Ad relevance questioned due to slow video player and loading issues.  
- User experienced freezing or failure to load video content completely.  
- Audio levels were too loud, causing discomfort during ad playback.

1. How relevant is this ad to you?

Video player was slow to load content Video content never loaded Ad froze or did not finish loading Video content did not start after ad Audio on ad was too loud Other issues

Is China using AI tokens mostly to produce micro-dramas?

You are about to exit Barron's official site.

By selecting the “Proceed” button below, you will be taken to an external website run by Hong Kong Tiimoot Information Technology Co., Limited (referred to as “HKT”), situated in Hong Kong. This site functions independently of Barron's, which does not have control over its content. It is important to review HKT's Privacy Statement, as their privacy policies may differ from ours. Please note that Barron’s is not accountable for HKT’s privacy policies or any other data management practices.

U.S. Treasury Aims to Curb Transactions Involving Chinese Biotech and AI

Dark, moody scene inside a quiet US Treasury corridor: on a polished mahogany table sits a sealed glass vial glowing eerie green with a spiraling DNA strand labeled “China,” beside it a matte‐gray AI processor chip pulsing with red warning lights. An outstretched silhouette hand of a Treasury official looms overhead in a firm, halting gesture. A folded US flag and a brass Treasury seal plaque rest at the edge of the table, casting deep shadows around the objects.

U.S. Treasury Wants to Rein In Deals for Chinese Biotech and AI - Barron's

- Users will be redirected to a third-party website owned by HKT.  
- Barron's does not control or endorse the content on the HKT website.  
- Privacy practices on HKT site are governed by its own Privacy Statement.  

You are now leaving Barron's website

By clicking on the “Proceed” button below, you will be redirected to a third-party website owned and operated by Hong Kong Tiimoot Information Technology Co., Limited. (“HKT”), which is located in Hong Kong. That website operates independently from Barron's and Barron's does not control the website. The privacy practices of HKT are subject to its Privacy Statement, so please read it closely. We are not responsible for HKT's privacy or other data-related practices.

U.S. Treasury Wants to Rein In Deals for Chinese Biotech and AI

The concept of using a single selfie to forecast the lifespan of a cancer patient appears to be a revolutionary advancement. While artificial intelligence (AI) has made significant strides in analyzing medical data patterns, its capability to consistently predict individual patient outcomes remains questionable. A clinical expert I consulted about the recent claims found the development intriguing, especially since medical practitioners have historically relied on facial observations to gauge health status. For instance, notes like “appears stated age” or “appears older than stated age” serve as indicators of overall health.

Researchers at Mass General Brigham in Massachusetts have expanded on this notion. They created an AI model called FAHR-Face, which is trained on over 40 million facial images to evaluate biological age and forecast survival rates in cancer patients. The biological age component, FAHR-FaceAge, estimates age by analyzing specific facial characteristics such as forehead lines, wrinkles between the nose and mouth, and temple structure. For the analysis of cancer survival, they developed FAHR-FaceSurvival, which utilized nearly 35,000 images of cancer patients. This model fixated on areas of the face, particularly below the eyes and along the bridge of the nose, that have historically been linked to mortality signals—echoing Hippocrates’ observations from 400 BC regarding signs of muscle deterioration and increased risk of death: “the nose sharp, the eyes sunken, the temples fallen in.”

The findings, recently published in The Lancet Digital Health, are compelling but not definitive. The AI-derived facial age showed a significant correlation with survival outcomes across 18 out of 28 cancer types. This implies effectiveness in predicting prognosis for some cancers but not universally applicable across all. Furthermore, Prof Miroslaw Bober highlighted a critical limitation: the tool requires validation against physicians' assessments and across diverse hospital settings and patient demographics. Notably, the facial-age measure did not significantly correlate with survival chances for Black patients, raising concerns about potential racial bias inherent in AI models due to data training processes and foundational assumptions. This bias has previously been observed in AI assessments of cardiac scans, used for heart disease diagnosis.

It’s essential to examine the implications of this study. The models were developed using retrospective data, identifying statistical relationships among thousands of patients, and extrapolating these to future survival predictions is a considerable leap. While a selfie might reveal some aspects of health, accurately forecasting future outcomes based on a single image is a complex endeavor, and more so, demonstrating its clinical utility remains a challenge.

Additionally, the reliability of selfies is a point of concern. Many have experienced how real-life appearances diverge from selfies, influenced by lighting, camera angles, expressions, or even the time of day. Cosmetic procedures like Botox or fillers, as well as treatments that alter facial contouring, complicate the relationship between facial appearance and health. The researchers acknowledged testing their age model against various photo conditions, yet, given the specific aging indicators they investigated, significant discrepancies in AI predictions seem inevitable.

Despite these challenges, the research is promising. The innovative tool, alongside conventional assessments, could help refine the typical evaluations that healthcare professionals perform. The real inquiry is whether this technology would enhance daily clinical practices beyond current methodologies, which already leverage years of expertise.

If you're curious enough to visit the project’s website to discover your biological age—perhaps to show off to friends—be mindful that uploading your selfie means your data will contribute to their broader database, as indicated in their consent forms. The tool currently focuses solely on estimating biological age, but one should remain cautious about its implications for overall health.

As a parting note, the researchers identified three primary factors that significantly age a person's appearance: alcohol consumption, drug use, and smoking. While these results are not revolutionary, they do shed light on how physicians can often gauge patient health from facial features, making one wonder if AI analysis of 40 million faces was really necessary for such insights.

Is AI truly better than doctors at predicting cancer survival? I have my reservations | Devi Sridhar

A dimly lit hospital office at night: a lone doctor in a white coat stands before a glowing computer screen filled with algorithmic code, stylized digital brain icon and cancer survival curves, their arms folded in a skeptical posture. Medical charts and a stethoscope lie on a desk bathed in the monitor’s eerie glow. Dark, moody atmosphere reflects doubt about AI’s predictions.

Can AI really predict cancer survival better than a doctor? I have my doubts | Devi Sridhar - The Guardian

- Researchers developed FAHR-Face, an AI model predicting cancer survival from facial analysis.  
- The model shows significant survival associations for some cancer types, but not all.  
- Concerns arise over racial bias and the accuracy of predictions based on selfies.  

The idea that a single selfie could predict how long someone with cancer may live sounds like an incredible breakthrough. AI is getting better at spotting patterns in medical data, but can it be used to reliably predict what will happen to an individual patient? I asked a clinical colleague what she thought about the recent headlines. She found it an interesting development given that doctors have long used facial appearance to assess patient health. For example, she noted that they may write “appears stated age” or “appears older than stated age” as a proxy for general health.

Now, researchers from Mass General Brigham, a healthcare institution in Massachusetts, have taken this idea a step further. They have developed an AI model trained on more than 40 million facial images – called FAHR-Face – to assess biological age and predict survival in cancer patients. For biological age, they developed FAHR-FaceAge, which estimates age from certain facial features such as the forehead, the wrinkles between the nose and mouth and the temples. For cancer survival, they developed FAHR-FaceSurvival, trained using almost 35,000 photos of cancer patients. This model focused more closely on the area below the eyes and the bridge of the nose – regions of the face “linked with dying”. Dr Raymond Mak, one of the study’s authors, noted that these parts of the face are the same as those referenced by Hippocrates in 400BC as signs of muscle wasting and mortality risk: “the nose sharp, the eyes sunken, the temples fallen in”.

The results, published in The Lancet Digital Health this month, are intriguing – but far from conclusive. Researchers found that AI-estimated facial age was significantly associated with survival in 18 of 28 cancer categories. So, basically, it works for some cancers, but not for others. In addition, Prof Miroslaw Bober noted, “The tool still needs testing against doctors’ own judgments and across different hospitals and patient groups. For example, the facial-age measure was not significantly associated with survival for Black patients after adjustment for other factors.” That finding adds to the growing concern that AI models can exhibit racial bias based on the data used to train them and underlying assumptions. For example, racial bias has been shown when AI models analyse cardiac magnetic resonance (CMR) scans – used to diagnose and monitor heart disease.

Let’s go back to what the published study does and doesn’t tell us. These models are built on retrospective data looking at statistical associations across thousands of patients, and it’s a jump to then accurately predict future survival from this. A selfie may contain clues about a patient’s health, but that is a long way from being able to predict their future – and a longer way still from proving that the prediction is useful in clinical practice.

I also worry about how unreliable selfies are. Just think of the times you’ve met people in real life and compared them with their profile photos. Anyone who has tried one of the online AI “guess your age” tools will know how much the result varies with lighting, camera angle, facial expression or even time of day. And then there’s botox, fillers, facelifts and even GLP-1 drugs, which alter facial fat and shape. How much can a single photograph really tell us about someone’s underlying health?

skip past newsletter promotion Free newsletter | Weekly Sign up to Matters of Opinion Guardian columnists and writers on what they’ve been debating, thinking about, reading, and more Enter your email Sign up after newsletter promotion

As anti-ageing treatments become the norm, visible age and underlying health will become an increasingly complex relationship. A face in a selfie that looks younger may not be necessarily healthier. The researchers do note that they tested their age model against lighting, expression and cosmetic surgery – but if predictions are based on specific signs of ageing in parts of the face that those all affect, I can’t see how it would be possible to avoid a certain degree of error in AI estimates.

None of this means the research isn’t promising. This new tool and the research team behind it are impressive. Alongside other tests that a doctor may perform to assess a patient’s health and future prognosis, it could take some of the guesswork out of “looks stated age”. But the question is whether it’s actually useful in daily practice beyond what clinicians are already doing when using experience and years of training.

If you’re tempted to go to the project’s public website to find out your biological age, even just to brag to friends about how young it says you look, be aware that by uploading your selfie, you will become part of the study too, as their consent forms indicate. Your face is not only being assessed, it’s contributing to the data used to develop the technology. Currently, the tool will only tell you your biological age, but I would be sceptical of what that may tell you about your overall health.

Finally, a pop quiz (I’m in the middle of the teaching semester): what top three things did the researchers find made people’s face age? Alcohol, drugs and smoking. No surprise there, but possibly an explanation as to why doctors can broadly read someone’s health status from their face. Something that perhaps we didn’t need AI scanning 40m faces to tell us.

Can AI really predict cancer survival better than a doctor? I have my doubts | Devi Sridhar

The Japanese yen surged to its highest point in over a month on Friday, driven by heightened safe-haven demand amid growing risk aversion that also resulted in a sharp decline in U.S. Treasury yields. This shift in the market pressure pushed the dollar down to its lowest value in nearly three weeks.

This currency fluctuation can be attributed to apprehensions surrounding the sustainability of the artificial intelligence (AI) spending wave, despite Brent crude oil prices remaining above $110 per barrel. This indicates that concerns regarding economic growth are currently overshadowing inflation issues.

The yen increased by 0.3% to 156.28 per dollar, marking it as the leading gainer among major currencies for the day. It is also on track for a 1% weekly increase, reflecting its most significant rise since early September.

The dollar index, a gauge of the greenback's performance against a group of currencies, was recorded at 101.52. It dipped to 101.44 on Thursday, marking its lowest point since the start of the month.

U.S. 10-year Treasury yields remained close to a one-month low at 4.953%, having decreased dramatically on Thursday amidst a sell-off in equities, triggered by renewed skepticism about the AI industry’s growth potential and enhancing the appeal of government bonds as a safe investment.

ING analysts noted the uncertainty surrounding whether yields will stay below 5% in the future. They anticipate that the 10-year yield could eventually rebound above this level, influenced by inflation, fiscal deficits, and AI developments.

In the current risk-averse climate, analysts indicated that this takes the pressure off the Federal Reserve to raise rates in their upcoming meeting. However, they advise against heavily investing in long-duration bonds for now.

Meanwhile, Brent crude prices surged over $114 per barrel, marking a 4% rise from the previous day after Houthi attacks in Saudi Arabia raised alarms about potential disruptions to regional energy supplies.

Amid rising safety demands, the favored currencies vary based on the associated risks. According to analysts at OCBC, uncertainties in the AI sector bolster the yen, while geopolitical tensions in the Middle East support both the Canadian dollar and the Norwegian crown due to escalating oil prices. Additionally, France's fiscal challenges have benefited the Swiss franc, highlighting how the U.S. dollar has become increasingly uncertain due to AI-related concerns.

On Thursday, the yen achieved its most substantial daily gain in nearly a month ahead of the upcoming Bank of Japan policy meeting. New data indicated that Japan's annual core consumer inflation surpassed the central bank's 2% target in September for the first time in nine months, suggesting broadening price pressures that may warrant a near-term interest rate adjustment. Most economists surveyed predict the bank will maintain its current rates next week, followed by a potential quarter-point increase in December.

The Canadian dollar also appreciated slightly, trading at over C$1.40 against the U.S. dollar, marking its strongest position in a month. The Swiss franc held steady at 0.8199 per dollar, close to its one-month high.

China's offshore yuan remained stable at 6.6715 per dollar, reaching its strongest level in over four years, while the Euro was nearly unchanged at $1.122.

Concerns regarding France's capability to manage its substantial debt are weighing on French government bonds and raising worries about other heavily indebted eurozone nations, impacting the euro negatively. The European Central Bank is anticipated to maintain its benchmark interest rate at 2.5% in the upcoming week.

The Australian dollar maintained its value at $0.7016, and the New Zealand dollar saw a slight increase of 0.1% to $0.5687. 

This report was authored by Li Gu in Shanghai and Ankur Banerjee in Singapore, with editing contributions from Kevin Buckland.

Yen gains strength as AI concerns drive investors toward safe havens.

A moody, dark trading floor lit only by a warm spotlight on a giant, gleaming golden yen coin at center. Shadowy trader silhouettes reach toward it in a frantic rush. Behind them, a cracked AI circuit–board mural fractures into shards. A single Japanese flag hangs overhead.

Yen shines as AI jitters spur safe-haven rush - Reuters

- The Japanese yen strengthened to its strongest level in over a month.  
- US Treasury yields fell sharply, pushing the dollar to a three-week low.  
- Concerns over AI spending and Middle East tensions impacted investor confidence.  

Summary Dollar pressured by sinking Treasury yields

Safe-haven yen and Swiss franc in demand

Investors fret about AI trade as Middle East tensions flare

SHANGHAI/SINGAPORE, Oct 23 (Reuters) - The Japanese yen hit its strongest level in more than a month on Friday, buoyed by safe-haven flows due to a fresh ​bout of risk aversion that also sent US Treasury yields sharply lower, pushing the dollar to a nearly three-week ‌trough.

The moves in the currency markets were driven by doubts over the AI spending boom even as Brent crude futures stay above $110 per barrel, suggesting concerns over economic growth outweighed inflation worries.

Sign up here.

The yen strengthened 0.3% on Friday to 156.28 per dollar, leading gains among major currencies. It was set for a 1% ​gain for the week, its strongest weekly rise since early September.

The dollar index , which measures the greenback against a basket of ​currencies, was at 101.52, after touching its lowest level since the beginning of the month at 101.44 ⁠on Thursday.

US 10-year Treasury yields hovered near a one-month low at 4.953%, after tumbling on Thursday when equities sold off on renewed concerns ​about the sustainability of the AI boom, burnishing the safe-haven appeal of sovereign debt.

"Whether yields stay below 5% is another question," said ING analysts ​said, adding they still expect the 10-year yield to snap back above 5% eventually, given inflation, fiscal deficit and AI trends.

"The risk-off environment takes pressure off the Fed to hike next week. But even given that, we'd not be loading up on duration here, at least not just yet," the analysts said.

Brent ​futures pushed above $114 per barrel, extending a 4% climb from the previous session after Houthi attacks in Saudi Arabia stoked concerns of further disruption ​to energy supplies from the region.

RISING DEMAND FOR SAFETY

While demand for safety is intensifying, the beneficiaries differ depending on the risk, OCBC analysts said.

AI jitters ‌favour the ⁠yen, Middle East tensions support the Canadian dollar and Norwegian crown due to rising oil prices, while French fiscal stress benefits the Swiss franc, they said, adding that for the US dollar, AI had become an increasingly two-way macro risk.

The Japanese yen posted its strongest daily gain in almost a month on Thursday, ahead of the Bank of Japan's policy meeting next week.

Data on Friday showed, Japan's annual core ​consumer inflation exceeded the central bank's 2% ​target in September for the ⁠first time in nine months, a sign of broadening price pressure that bolsters the case for a near-term interest rate hike. The majority of economists polled by Reuters except a hold on Friday of next week, ​followed by a quarter-point hike at the following meeting in December.

The Canadian dollar firmed past C$1.40 ​against the greenback, ⁠up 0.1% to its strongest level in a month. The Swiss franc was steady at 0.8199 per dollar, near its firmest level in a month.

China's offshore yuan was little changed at 6.6715 per dollar, its strongest level in more than four years. The Euro was largely flat at $1.122.

Worries about ⁠France's ability ​to tackle its huge debt burden have hurt French government bonds and raised ​concern about other heavily indebted euro-area countries, weighing on the euro.

The European Central Bank is expected to hold its benchmark rate at 2.5% next week.

The Australian dollar held steady ​at $0.7016, while the Kiwi dollar edged up 0.1% to $0.5687.

Reporting by Li Gu in Shanghai and Ankur Banerjee in Singapore; Editing by Kevin Buckland

Our Standards: The Thomson Reuters Trust Principles. , opens new tab

Yen shines as AI jitters spur safe-haven rush

The government of Togo is encouraging its citizens to contribute texts, audio recordings, and translations to create artificial intelligence models that reflect the country's rich linguistic diversity, as reported by the AI for Developing Countries Forum.

This initiative is being viewed not merely as a localized experiment but as a template for how developing countries can leverage technology for cultural preservation and economic significance in the era of digital transformation.

Africa boasts approximately 2,000 languages, yet very few are inclusively represented in AI platforms, as highlighted by experts on The Conversation. Failing to integrate these languages into technological advancements carries the risk of alienating vast populations and widening the technological divide.

The challenge lies in AI's limited ability to accommodate a majority of African languages. This shortfall stems from the disparity between the presence of native speakers and the digital representation of their languages, according to African Business. Research published in the Proceedings of Machine Learning Research reveals that across a range of language models, only 41 African languages are represented, with a mere four—Amharic, Swahili, Afrikaans, and Malagasy—consistently included. More than 98% remain unsupported.

Despite a rapid increase in ChatGPT usage in Africa, as noted by The Economist, even the most advanced language models struggle with African languages, often performing 10 to 30 percentage points worse in performance assessments in comparison to English. Difficulties arise from the multitude of dialects and the phenomenon of "code-switching," where speakers intertwine different languages in conversation.

The core issue, as stated by experts, is a lack of data. Most language training for lesser-represented languages relies heavily on Bible texts, while high-quality audio for speech recognition is limited. To address this gap, researchers are actively recording natural speech patterns.

Moreover, the challenges extend beyond just data scarcity. Infrastructure inadequacies present further barriers, such as the absence of essential tools like dictionaries and glossaries. There are also shortcomings in the technology required to support these languages, including keyboard layouts, fonts, spell-checkers, and tokenisers, which all affect how languages are processed and understood. These factors create issues with spelling variations and tonal distinctions within dialects.

The dominance of texts in LLMs mirrors cultural assumptions prevalent in the Global North, emphasizing individualism and competition, as pointed out by University World News. This reinforces a historical pattern of epistemic marginalization experienced by the Global South.

The lack of representation of African languages in AI models has now become a pressing issue for policymakers and technology leaders across the continent, as stated by Semafor. The World Bank indicates that Africa accounts for just 0.6% of global data-center capacity, despite its significant share of the world’s population.

Currently, Africa’s primary connection to the AI industry has been through its mineral resources like cobalt and copper; however, the ultimate opportunity may lie in enhancing productivity and harnessing the full potential of AI. At present, the continent lacks the necessary infrastructure and energy resources to capitalize on this, leaving meaningful growth linked to AI unlikely.

African tech companies are urged to progress beyond being mere suppliers or consumers within the AI landscape. Instead, there is a call for the continent to be a leader in AI development where the most significant benefits can be realized.

Building AI models from scratch is often costly and resource-intensive. By adopting and tailoring existing models, Africa can accelerate its technological advancements. Togo’s initiative demonstrates that the future of AI in Africa may depend less on the scale of the models created and more on the ability to craft technology that serves the greatest number of people effectively.

The AI dilemma in Africa

A shadowy server room lit by a single amber spotlight reveals racks of half-powered machines and tangled cables; on one monitor, a flickering circuit-style map of Africa pulses weakly, while a lone adult African engineer stands in silhouette studying failing code streams, the scene heavy with deep shadows to illustrate the continent’s AI infrastructure struggle.

Africa’s AI challenge - The Week

- Togo is inviting citizen contributions to create AI models reflecting its linguistic diversity.  
- Over 98% of African languages remain unsupported in existing AI language models.  
- The lack of infrastructure and resources hinders Africa's ability to leverage AI effectively.

Togo has asked its citizens to contribute texts, recordings and translations to “develop AI models that address the linguistic diversity of the country”, said AI for Developing Countries Forum.

More than just a “regional experiment”, it is a “blueprint for how developing nations can use technology to assert their cultural sovereignty and economic relevance in the digital age”.

Africa has around 2,000 languages but very few are represented comprehensively in AI models, said academics on The Conversation. Failing to be a part of the tech revolution not only “marginalises millions of people” but also “increases the technology divide”.

The Week Escape your echo chamber. Get the facts behind the news, plus analysis from multiple perspectives. SUBSCRIBE & SAVE Sign up for The Week's Free Newsletters Join more than 350,000 subscribers and keep yourself informed with a selection of The Week’s most interesting, enlightening and entertaining stories - plus daily puzzles. Join more than 350,000 subscribers and keep yourself informed with a selection of The Week’s most interesting, enlightening and entertaining stories - plus daily puzzles. Sign up

Latest Videos From The Week US Watch full video here:

98% of African languages ‘unsupported’

AI’s inability to handle most African languages comes down to the “gap between the number of real-world speakers and a language’s digital footprint,” said African Business.

According to research published in the Proceedings of Machine Learning Research, across six large language models (LLMs), eight small language models and six specialised small language models, only 41 African languages were represented. Only four – Amharic, Swahili, Afrikaans and Malagasy – were “always handled”. “Over 98% of African languages went unsupported.”

According to OpenAI, the number of ChatGPT users in Africa is “growing faster than in any other region, said The Economist. Yet, despite the surge in popularity, even the best LLMs “falter in African languages, typically scoring 10-30 percentage points lower on benchmarking tests than they do in English”. The models mainly struggle with numerous dialects and “code-switching”, which is when speakers use different languages at the same time.

“The root problem is lack of data.” Much of the training for “smaller” languages is taken from Bible verses. High-quality audio fit for speech recognition is also “scarce”, so researchers are “recording how people talk” to build the data supply.

Sign up to get Today’s Best Articles in your inbox A free daily email with the day’s biggest news and analysis – and the best features from TheWeek.com Contact me with news and offers from other Future brands Receive email from us on behalf of our trusted partners or sponsors

It’s not just a lack of text and speech that is the problem, said the academics on The Conversation, it’s also the infrastructure that supports it. “Basic tools” like dictionaries, terminologies and glossaries are missing. Then there are inadequacies in African language keyboards, fonts, spell-checkers and tokenisers – which “break text into smaller pieces so a language model can understand it”. All of these impact orthographic variation – how the same words are spelled differently across regions – tone marking and “rich dialect diversity”.

‘Global North cultural assumptions’

The texts that dominate LLMs “reflect Global North cultural assumptions, about individualism, competition and about globalisation”, said University World News. By definition, therefore, this is a continuation of the “long history of epistemic marginalisation” of the Global South.

The lack of African languages in LLMs has become a “major concern for the continent’s policymakers and tech leaders”, said Semafor. The World Bank estimates that Africa has only 0.6% of global data-centre capacity, despite having 18% of the world’s population.

Africa’s “main link” to the AI revolution has been through minerals such as cobalt and copper, but the “bigger prize” is through productivity, or being able to harness the end product of AI. Currently, however, there isn’t the infrastructure or energy to achieve this, meaning that AI is “unlikely to add meaningfully to growth”.

African AI companies need to break beyond being “suppliers to the AI revolution or just consumers of finished AI products”. Their continent needs to be the place where the “revolution delivers its biggest gains”.

Building AI models from scratch is “expensive” and uses “limited resources”. By adopting and adapting existing models, the continent’s outputs can evolve more quickly. And Togo’s initiative suggests that the future of AI in Africa “may hinge less on who builds the biggest model than on who makes the technology useful to the most people”.

Africa’s AI challenge

A federal appeals court has overturned a prior decision that prohibited Mississippi's law prohibiting diversity, equity, and inclusion (DEI) programs in public schools, transferring the case to a different district judge amid concerns over judicial integrity stemming from AI use. The 5th U.S. Circuit Court of Appeals, based in New Orleans, invalidated the preliminary injunction against the law, which had previously been issued by U.S. District Judge Henry Wingate. This ruling is seen as a win for Mississippi Attorney General Lynn Fitch's office, which criticized Judge Wingate for relying on generative AI tools that resulted in numerous inaccuracies.

Judge Jerry Smith, who wrote for the three-judge panel, criticized the lower court's handling of the case, describing Wingate's ruling as a "botch job" filled with errors and inconsistent facts not supported by the legal record. Smith remarked that the lower court appeared to manipulate the governing law to align with a specific policy agenda.

The appeals court provided the incoming judge with clear guidelines for evaluating the law, emphasizing that it does not extend to the students and highlighting the importance of balancing the state’s regulatory authority over education with the free speech rights of teachers and professors. Robert McDuff, representing the plaintiffs from the Mississippi Center for Justice, acknowledged that while the 5th Circuit's decision alleviated some of their concerns regarding free speech in educational settings, they would continue to challenge other aspects of the law.

Despite requests for comments from Fitch's office and Judge Wingate's chamber, there was no immediate response. All judges involved, including Wingate, are Republican appointees. The legal environment has become increasingly complex as courts grapple with inaccuracies stemming from AI's involvement in legal documentation, raising questions about its implications for the judicial process.

The case originates from a challenge by various educators and advocacy organizations against a law signed by Mississippi's Republican leadership last year, which restricts public educational institutions from requiring DEI training. The plaintiffs contended that this law violates the First and Fourteenth Amendments of the U.S. Constitution and succeeded in persuading Wingate to impose an initial temporary restraining order to suspend the law’s implementation.

However, that order was later found to include fabrications and spurious legal citations, which caught the attention of Senator Chuck Grassley, the chairman of the Senate Judiciary Committee. In response to Grassley, Wingate attributed the discrepancies to a law clerk's unreviewed implementation of the AI tool Perplexity. Following these issues, Wingate issued a preliminary injunction to maintain the block on the law, despite the emerging concerns about AI errors undermining the legal reasoning.

Appealing to the 5th Circuit, Mississippi's legal team argued that the reliance on AI compromises the court's foundational legal arguments, advocating for the reassignment of the case. Conversely, the attorneys for the plaintiffs urged the court to leave the case with Wingate, framing the AI use as a regrettable oversight that did not diminish the constitutional issues at play. Nevertheless, Smith determined that the high number of errors indicated a necessity for reassignment, cautioning against Wingate’s potential biases in future proceedings.

The case is officially referenced as Jackson Federation of Teachers, et al., v. Fitch, in the 5th U.S. Circuit Court of Appeals.

Appeals court removes restriction on Mississippi's DEI ban and dismisses judge due to AI mistakes

A moody, dimly lit courtroom interior: a heavy marble bench where a massive stone slab engraved “DEI” is being hoisted away by unseen forces, dust swirling in the air; to the side, a lone judge in a dark robe turns and exits through an ornate doorway, his gavel dropping to the floor; nearby, a cracked tablet lies on the ground, its screen flickering with corrupted AI code; the Mississippi state flag hangs quietly on the wall as natural light streams through high windows, casting long, dramatic shadows.

Appeals court lifts block on Mississippi's DEI ban, ejects judge over AI errors - Reuters

- The 5th Circuit vacated an injunction against Mississippi's ban on DEI in public schools.  
- A lower court's AI-generated errors led to reassignment of the case to a new judge.  
- The ruling clarified that discussions on race and civil rights can continue in schools.  

Summary 5th Circuit throws out preliminary injunction against Mississippi law barring DEI mandates in public schools

Lower-court judge said law clerk's unvetted use of AI caused hallucinations in court order

Panel says case goes to different district judge for further review

Oct 22 (Reuters) - A federal appeals court on Thursday vacated a ruling that ​blocked Mississippi's ban on diversity, equity and inclusion programs in public schools, reassigning the case away from a district court judge who had acknowledged that a key ‌ruling contained AI-generated errors.

The New Orleans-based 5th US Circuit Court of Appeals threw out , opens new tab the preliminary injunction against the state law and reassigned the civil rights lawsuit to a different district judge for further review.

Sign up here.

The decision is a victory for Republican Mississippi Attorney General Lynn Fitch's office, which had argued that US District Judge Henry Wingate's apparent misuse of generative AI tools had compromised the integrity of the judicial process.

US Circuit Judge Jerry Smith, writing for the three-judge ​panel, took Wingate to task for a "botch job" that led to a ruling that contained "sloppy edicts" and "several factual assertions nowhere to be found in the record" and got the ​law wrong.

"The district court — or perhaps, being charitable, its AI agent — ignored and manipulated governing law, apparently to satisfy a policy agenda," Smith wrote.

The appeals ⁠court laid out detailed instructions for how the new judge should assess the law.

It made clear it does not apply to students, and it said that while states have broad power ​to regulate their own curricula and publicly funded speech, the new judge will need to balance the state's interests against the individual speech rights of teachers and professors.

Robert McDuff, a lawyer for the ​plaintiffs at the Mississippi Center for Justice, in a statement said the ruling addressed some of their concerns by making it clear that public school teachers and students could still discuss issues related to race, sex, and civil rights on a variety of fronts.

"The 5th Circuit has sent the case back for further review before a new trial court judge and we will be addressing other aspects of this law going forward, but this ruling eliminates ​one of our major concerns about it," he said.

A spokesperson for Fitch did not respond to a request for comment, nor did Wingate's office. Wingate and the judges on the appeals court panel are all Republican appointees.

State and ​federal courts have grappled with a wave of AI-generated inaccuracies in filings . The cases have led to mounting concerns about the technology's role in the legal process, including whether and how judges are using AI programs to ‌help craft opinions.

US ⁠courts do not have uniform rules about lawyers' use of AI, and the policies can differ from one judge to another about their own use of the technology.

The case before the 5th Circuit concerned a challenge by teachers, professors, students and advocacy groups to a law Republican-led Mississippi enacted last year that bars government-run educational institutions from mandating activities like diversity training or promoting DEI.

The plaintiffs argued the law violated the US Constitution's First and Fourteenth Amendments, and convinced Wingate to issue a temporary restraining order blocking its implementation.

But that order was later found to contain fabricated names and nonexistent legal authorities, ​prompting scrutiny from Republican Senator Chuck Grassley, the ​current chairman of the Senate Judiciary Committee.

Wingate in ⁠a subsequent letter told Grassley that the errors stemmed from a law clerk's unvetted use of the AI platform Perplexity. Even after the errors in that initial order, Wingate continued to block the law by issuing a preliminary injunction.

Lawyers for Mississippi argued on appeal that even after Wingate replaced the faulty ​ruling, the AI-related errors cast doubt on the court's core legal reasoning, and they urged the appeals court to reassign the case.

Attorneys representing the ​teachers and civil rights ⁠advocates challenging the law had asked the appeals court to keep the case with Wingate, arguing that the AI use was a regrettable "shortcut" that did not affect the ultimate constitutionality of the state's DEI restrictions.

But Smith said the case "easily calls for reassignment." He said given the number of errors in Wingate's ruling, the judge might be tempted in future proceedings to want to unjustly find a way "to prove that, even with ⁠the correct law ​and fact, that his initial view should prevail."

The case is Jackson Federation of Teachers, et al, v Fitch, 5th ​US Circuit Court of Appeals, No. 25-60496.

For the plaintiffs: Robert McDuff of Mississippi Center for Justice

For the defendant: Anthony Shults of the Mississippi Attorney General's Office

Read more:

Reporting by Mike Scarcella

Our Standards: The Thomson Reuters Trust Principles. , opens new tab

X

Facebook

Linkedin

Email

Link Purchase Licensing Rights

Appeals court lifts block on Mississippi's DEI ban, ejects judge over AI errors

U.S. stock markets faced a significant downturn today, with the Nasdaq experiencing its most considerable decline since early June, largely driven by concerns surrounding artificial intelligence (AI) investments following several notable earnings reports. 

The S&P 500 closed well into negative territory, while the tech-heavy Nasdaq dropped 2.6%, primarily impacted by losses in the semiconductor sector after a series of high-profile AI-related earnings announcements. The Dow Jones, however, managed to gain slightly right before the markets closed.

Meta Platforms exceeded analyst expectations with its recent results, yet its stock fell by 7.0%. The parent company of Facebook recorded its first cash outflow in over a decade, attributed to investments in AI, and revealed plans for a non-voting share issuance in a two-for-one split to raise additional funds for these ventures.

Tesla saw a decline of 3.1%, despite reporting an unexpected positive cash flow and reducing its capital expenditure projections for 2026.

The Philadelphia Semiconductor Index, a key contributor to the recent bull market, plunged by 6.7%, marking its steepest one-day drop since late June. Following this decline, the SOX index is now down more than 26% from its record high on June 22 but boasts an impressive 51.5% increase year-to-date. In comparison, the S&P 500 has risen 11.9% during the same timeframe.

Paul Nolte, a senior wealth adviser at Murphy & Sylvest, commented, “It's a wonderful day on Wall Street. The technology sector, once the darling of the markets, seems to be struggling.” He noted that investors are starting to question the sustainability of the AI narrative, which is impacting heavily tech-weighted indices.

On the global front, tensions escalated as Iran-backed Houthis launched attacks on Saudi Arabia, pushing crude oil prices to a five-month peak. This geopolitical instability has led to investor concerns about persistently high inflation, possibly prompting the U.S. Federal Reserve to continue raising interest rates.

Labor Department data indicated that initial unemployment claims remain near multi-decade lows. This underlying stability in the labor market appears to give the Fed the leeway needed to address inflation, with financial markets anticipating a possible 25 basis point hike in the policy interest rate before the end of the year.

In today’s trading, the Dow Jones Industrial Average increased slightly by 11.89 points, or 0.02%, to 52,867.98, while the S&P 500 fell by 93.74 points, or 1.21%, settling at 7,663.43. The Nasdaq Composite lost 694.00 points, or 2.60%, closing at 26,008.86.

As the third-quarter earnings season progresses, 111 companies from the S&P 500 have reported their results, with about 83% surpassing analyst expectations, according to the latest data from LSEG.

Procter & Gamble highlighted the strain of rising costs on profits but managed to exceed forecasts, leading to a 5.6% stock increase, which buoyed the consumer staples sector. Blackstone rose by 8.4% after indicating that widening credit spreads on AI-related debts could yield higher returns. Lockheed Martin's stock climbed 7.0% following an upward revision of its annual sales and profit expectations.

Spotify surged by 11.1% after projecting a stronger-than-expected fourth-quarter profit, crediting its AI-driven features for anticipated user growth. Union Pacific gained 3.6% following an earnings surprise supported by core pricing improvements, despite noting the adverse impact of rising fuel costs.

In the broader semiconductor market, key players like Intel, AMD, SanDisk, and Marvell saw declines exceeding 8%. 

On the NYSE, advancing stocks slightly outnumbered declining ones with a ratio of 1.01-to-1. There were 173 new highs and 366 new lows. Meanwhile, on the Nasdaq, 2,022 stocks advanced, while 2,766 decreased, leading to a 1.37-to-1 ratio of declines to advances.

The S&P 500 recorded 40 new 52-week highs alongside four new lows, while the Nasdaq Composite logged 82 new highs and 342 new lows.

Overall trading volume across U.S. exchanges reached 18.79 billion shares, surpassing the 17.19 billion average for the previous 20 trading days.

Nasdaq drops 2.6% amid concerns over AI spending triggering a tech selloff

A dimly lit trading floor cloaked in deep shadows, rows of empty workstations bathed in the glow of red numbers, a giant digital screen dominating the scene displaying “NASDAQ –2.6%” with a sharp, downward-sloping arrow; a lone trader’s silhouette stares up in tense disbelief as price tickers on smaller monitors flicker with plunging tech stock symbols.

Nasdaq plunges 2.6% as AI spending fears spark tech selloff - Honolulu Star-Advertiser

- U.S. stocks fell sharply, with the Nasdaq dropping 2.6% amid AI earnings fears.  
- Meta reported its first cash outflow in a decade, causing a 7% stock drop.  
- Semiconductor Index plunged 6.7%, highlighting concerns about the technology sector's stability.  

NEW YORK >> U.S. stocks tumbled today and the Nasdaq posted its steepest daily drop since early June as a batch of high-profile earnings reports ignited fresh fears about the AI spending race.

The S&P 500 ended deep in negative territory. But the tech-laden Nasdaq, dragged down by microchips, suffered a steeper loss, falling 2.6% in the wake of a spate of high-profile AI-related earnings results.

The blue-chip Dow eked out a nominal gain minutes before the closing bell.

Meta Platforms’ results beat analyst expectations, but its stock slid 7.0% as the Facebook parent posted its first cash outflow in more than a decade due to its AI-related expenditures and announced plans to issue non-voting shares in a two-for-one split to raise capital to spend on the nascent technology.

Tesla shares fell 3.1% even though the electric car maker unexpectedly posted positive free cash flow in the quarter and cut its 2026 capital spending plans.

The Philadelphia SE Semiconductor Index, which has driven much of the current bull market’s gains, plunged 6.7%, its biggest daily percentage drop since June 23.

Don't miss out on what's happening! Stay in touch with breaking news, as it happens, conveniently in your email inbox. It's FREE! Email Sign Up By clicking to sign up, you agree to Star-Advertiser's and Google's Terms of Service and Privacy Policy. This form is protected by reCAPTCHA.

With Thursday’s decline, the SOX is off more than 26% from its record closing high reached on June 22, yet it remains up 51.5% so far this year. The S&P 500 has risen 11.9% over the same period.

“It’s a wonderful day on Wall Street,” said Paul Nolte, senior wealth adviser & market strategist at Murphy & Sylvest in Elmhurst, Illinois. “The only part of the market that is really suffering here is technology, which has been the golden child of the markets up until about June.”

“Investors are starting to realize that there are some holes in the AI story,” Nolte added. “And that’s having an impact on the indices because they’re so heavily weighted toward technology.”

On the geopolitical stage, Iran-backed Houthis launched new strikes against Saudi targets, sending crude prices to a five-month high. Investors worried that ongoing Middle East strife could keep inflation hot and force the U.S. Federal Reserve to keep raising interest rates.

Data from the Labor Department showed initial claims for unemployment benefits kept drifting along near multi-decade lows. Signs of labor market stability will allow the Fed to focus on fighting inflation. Financial markets are currently pricing in the probability that the central bank will raise its policy rate by another 25 basis points before year-end.

The Dow Jones Industrial Average rose 11.89 points, or 0.02%, to 52,867.98, the S&P 500 lost 93.74 points, or 1.21%, to 7,663.43 and the Nasdaq Composite lost 694.00 points, or 2.60%, to 26,008.86.

Third-quarter earnings season is well underway, with 111 of the companies in the S&P 500 having reported. Of those, 83% have beaten analyst expectations, according to the most recent LSEG data.

Procter & Gamble warned that rising cost pressures would take an increasing bite out of profit, but beat analyst estimates, sending the stock 5.6% higher and supporting consumer staples.

Blackstone advanced 8.4%, after the company said widening credit spreads on AI-related debt could present an opportunity to lock in higher returns.

Lockheed Martin climbed 7.0% after the aerospace/defense firm raised its annual sales and profit forecast.

Spotify rose 11.1% after the Swedish music streaming firm forecast fourth-quarter profit above Wall Street estimates, betting on its AI-powered features to drive user growth.

Union Pacific gained 3.6% following the company’s quarterly earnings beat, which was helped by core pricing gains, although the railroad flagged a hit from rising fuel prices.

Chipmakers Intel, AMD, Sandisk and Marvell all tumbled more than 8% amid the broader semiconductor selloff.

Advancing issues outnumbered decliners by a 1.01-to-1 ratio on the NYSE. There were 173 new highs and 366 new lows on the NYSE.

On the Nasdaq, 2,022 stocks rose and 2,766 fell as declining issues outnumbered advancers by a 1.37-to-1 ratio.

The S&P 500 posted 40 new 52-week highs and four new lows while the Nasdaq Composite recorded 82 new highs and 342 new lows.

Volume on U.S. exchanges was 18.79 billion shares, compared with the 17.19 billion average for the full session over the last 20 trading days.

Additional reporting by Tharuniyaa Lakshmi and Ragini Mathur in Bengaluru.

Nasdaq plunges 2.6% as AI spending fears spark tech selloff

On October 22, the world reacted to alarming news: PHP, a coding language powering approximately 70% of all websites that many users may not even recognize, had been exposed with a significant security vulnerability. This discovery prompted urgent responses from businesses and government entities to address the issue.

The vulnerabilities are located within the foundational code of PHP. The National Security Agency (NSA) has raised concerns that cybercriminals could exploit these vulnerabilities, allowing them potential access to web servers to steal sensitive information, disrupt services, or navigate through networks that support everyday digital activities for the average user.

Experts emphasize that the primary responsibility for addressing these vulnerabilities lies with software vendors and the organizations that serve their customers. Nonetheless, there are fundamental measures individuals can adopt to enhance their security.

One of the key recommendations is to implement multi-factor authentication (MFA). This security process involves two verification methods for accessing secure sites, such as a password paired with a text message code or a security key. Emily Austin, a principal security researcher at Censys, advises using distinct passwords for different sites and keeping an eye on transaction statements. Austin notes, however, that in cases like the PHP vulnerabilities, individual users have limited recourse, relying mainly on the service providers to patch the vulnerabilities effectively.

These flaws in PHP are not isolated but reflect a pattern of serious software bugs becoming increasingly prevalent. The capabilities of modern large language models have made it easier to uncover and exploit previously hidden weaknesses in commonly used software.

To safeguard against potential exploitation, Dan Guido, CEO of Trail of Bits, advocates for enabling automatic updates across all devices, including smartphones, computers, and home routers. Security updates are often deployed by companies in response to known vulnerabilities, and staying up to date can significantly minimize risks from cyber threats.

However, Guido points out that users still depend on the diligence of service providers. “Even if a user adheres to best practices, like utilizing two-factor authentication, their data can still be compromised,” he warns. “You can’t update a hospital’s website.”

Runa Sandvik, a digital security expert and founder of Granitt, stresses the importance of keeping all applications and services updated, particularly for internet-connected devices. She cautions that many consumers might overlook devices that do not receive regular updates and urges checking with manufacturers for any available improvements.

Dr. Daniel Gardham, a lecturer at the University of Surrey's Centre for Cyber Security, highlights the necessity for organizations to identify the software they are utilizing, apply patches quickly, and investigate if they have suffered any breaches. He notes that AI advancements have given attackers a distinct advantage. “The speed at which vulnerabilities can be uncovered is increasing, but addressing them across diverse systems remains a considerable challenge,” he explains. “How effectively organizations navigate this challenge will be crucial in determining whether AI enhances or diminishes internet safety.”

In light of such cybersecurity threats, Nic Adams, CEO and co-founder of 0rcus, cautions that the emergence of significant vulnerabilities often leads to increased activity from scammers and cybercriminals. He advises users to be vigilant about unsolicited links and to verify any urgent messages related to securing accounts by directly accessing the company’s app or website rather than through provided links.

On a more reassuring note, banks have been preparing for such vulnerabilities and have worked diligently to eliminate obvious targets for cybercriminals, according to Paul Benda, executive vice president of risk, fraud, and cybersecurity at the American Bankers Association. Despite this preparation, Benda acknowledges the evolving nature of cyber threats and the importance of consumer vigilance, particularly when shopping at less secure websites compared to trusted banking platforms.

He advises consumers against reusing passwords on shopping sites and encourages the adoption of passkeys whenever possible, as these provide a digital solution that enhances security and mitigates phishing risks.

Overall, as organizations and users navigate the implications of these recent PHP vulnerabilities, remaining informed and proactive in cybersecurity practices is more crucial than ever.

Five strategies for staying safe online amid AI-enhanced hacking threats

In soft morning sunlight, an adult at a wooden desk types on a laptop as dark, tentacle-like streams of binary code lunge from the screen. A radiant padlock-shaped shield hovers over the keyboard, deflecting the tendrils. Above the laptop, five glowing icons – a lock, fingerprint, update gear, crossed-out email hook, and secure Wi-Fi symbol – float in an arch, each marked with a green check. The scene evokes confidence and vigilance in online safety.

Five ways to keep safe online in the wake of AI-supercharged hacking - Reuters

- PHP, used by 70% of websites, has a serious vulnerability prompting urgent fixes.  
- Experts emphasize organizations' responsibility to patch vulnerabilities and protect users' data.  
- Consumers should enable multi-factor authentication and be cautious of phishing scams.

Oct 22 (Reuters) - Much of the world woke Wednesday to news that PHP, a coding language most people have never heard of but that underlies roughly 70% of all websites, had a hackable vulnerability, sending businesses and governments scrambling to fix it.

While the PHP glitches sit in the underlying code itself, the NSA ​warned , opens new tab that attackers could use them to take control of web servers to steal data, disrupt services or move through networks on services the ‌average person relies on every day.

Sign up here.

Experts say responsibility mostly lies with software providers and organizations that serve customers. But there are basic steps everyone should take to protect themselves.

USE MULTI-FACTOR AUTHENTIFICATION

MFA, or multi-factor authentication, requires two forms of verification to access a protected site, such as a password plus a text-based code or security key. Emily Austin, principal security researcher at cybersecurity firm Censys, recommended unique passwords for individual ​websites, and for people to monitor any transactions.

“But realistically, for situations like the PHP vulnerabilities, there really isn’t much an end user can do beyond hope ​the services they rely on either don’t use PHP or have people actively working on patching,” Austin said.

The PHP flaws are the ⁠latest in a series of severe software bugs made worse by increasingly capable and accessible large language models. Long-hidden flaws in widely used software, once too obscure or complex ​for humans to exploit, are now much easier to surface and weaponize , opens new tab

MAKE HARDWARE UPDATES AUTOMATIC

Dan Guido, CEO and co-founder of cybersecurity firm Trail of Bits, recommended people “turn on automatic updates for ​your phone, computer, and home router.” Companies typically issue security updates when patches are available for known vulnerabilities, so installing them promptly limits the window for bad actors.

Even so, consumers remain at the mercy of the services they rely on, Guido said.

“Someone visiting a website can do everything right, like use two-factor authentication, and still have their information exposed,” Guido said. “You can't patch your hospital's website.”

Most of the ​work falls on companies running services, Guido said. “They need to deploy fixes faster and limit the damage that compromised systems can cause.”

Item 1 of 3 The words "AI Hacking" and PHP code are displayed on a screen in this illustration taken on October 22, 2026. REUTERS/Dado Ruvic/Illustration [1/3] The words "AI Hacking" and PHP code are displayed on a screen in this illustration taken on October 22, 2026. REUTERS/Dado Ruvic/Illustration Purchase Licensing Rights , opens new tab

Runa Sandvik, the founder of Granitt and a ​digital security expert, said people need to update apps and services on their devices whenever they're available, and to not forget about internet-connected devices around the house.

"I can see a scenario in which ‌some people ⁠have some sort of IoT device at home that maybe doesn't get frequent updates, maybe doesn't get any updates at all," Sandvik said. "It's worth double checking with the vendor to see if there is anything available."

PATCH, PATCH, PATCH

Dr. Daniel Gardham, a lecturer at the Surrey Centre for Cyber Security and the University of Surrey, said organizations must know what software they run, patch promptly and check whether they've already been compromised. In the AI age, attackers have the advantage, Gardham added.

“Finding vulnerabilities is becoming faster and easier, but fixing them ​across the range of systems that rely ​on this software is much harder,” Gardham ⁠said. “How well organizations manage that gap will determine whether AI makes the internet safer or more vulnerable.”

BEWARE OF SCAMMERS

Any time a major cybersecurity threat makes the news, scammers and cybercriminals will try to take advantage, said Nic Adams, co-founder and CEO of 0rcus, a ​cyberkinetics company.

Be wary of unsolicited links. "When these headlines hit, check the affected company's instructions by opening its app or typing its website ​address yourself," Adams said. "An ⁠urgent message telling you to 'secure your account' deserves extra scrutiny."

YOUR MONEY IS SAFE

Banks have been preparing for this moment since frontier models came out, and have "worked extremely hard to ensure there is no more low-hanging fruit for the bad guys to go after," said Paul Benda, executive vice president of risk, fraud and cybersecurity at the American Bankers Association.

However, Benda added, "we are in ⁠a new normal" ​in terms of cyber threats, and consumers need to be aware, especially in the wake of the ​PHP vulnerabilities, when shopping with random retailers or websites "where their security is not going to be as good as your bank."

Benda said consumers should not reuse passwords across shopping sites, and should consider using passkeys , opens new tab where possible, which are ​digital keys specific to certain websites, that avoid the need to enter a password into a website and a phishing resistant.

(This story has been refiled to add the dropped words 'unsolicitated links', in paragraph 20.)

Reporting by AJ Vicens in Detroit; Editing by Jonathan Spicer

Our Standards: The Thomson Reuters Trust Principles. , opens new tab

X

Facebook

Linkedin

Email

Link Purchase Licensing Rights

Five ways to keep safe online in the wake of AI-supercharged hacking

On Thursday, the AI sector's latest turbulence weighed down global stock markets, while escalating oil prices reached their highest point since May, driven by concerns over the conflict involving Iran.

The S&P 500 saw a 1.2% decline, marking its most significant drop since July, while the Nasdaq composite plunged by 2.6%. The impact of this tech stock weakness was so pronounced that it overshadowed positive performances among other sectors of the U.S. market, where most stocks appreciated. In contrast, the Dow Jones Industrial Average managed a modest increase, rising 11 points or nearly 0.1%.

Investors are increasingly anxious about AI stocks amid fears that the booming growth may face challenges due to rising borrowing costs for constructing data centers and acquiring additional AI chips. Additionally, there are concerns that stock prices have soared excessively following a prolonged market frenzy, despite ongoing growth in the sector.

Lam Research, which specializes in equipment and services for semiconductor manufacturing, reported better-than-expected profits and revenues for the summer quarter. However, instead of rising, its stock plummeted by 11.8%.

Similarly, other major tech companies faced declines following their quarterly announcements, which included some positive aspects. Meta Platforms, known for its Muse AI initiative, dropped 7% despite exceeding revenue expectations for the quarter, losing significant market value as profits did not meet projections. The parent company of Facebook and Instagram also revealed plans for a new stock class aimed at funding AI projects and other ventures.

Tesla's stock fell 3.1% after it posted a quarterly profit that surpassed analyst expectations, although its revenue fell short of forecasts.

The broader decline in AI stocks had a noticeable effect on Nvidia, which dropped 4.4%. As the largest company by market capitalization on Wall Street, Nvidia's stock movements significantly influence the S&P 500, making it a primary factor in pulling the index lower.

CoreWeave, which has financed purchases of Nvidia's AI chips for cloud rental services, has experienced volatility due to concerns regarding its debt levels, leading to a 10.1% decrease in its stock.

The downturn for AI and tech stocks began in Asia, impacting numerous indexes populated by industry players. The South Korean Kospi fell by 3.5%, Hong Kong’s Hang Seng dipped 3%, and Japan's Nikkei 225 dropped by 2.7%.

Additionally, oil prices surged, contributing to the declines in global stock markets. Brent crude oil prices jumped by 4% to $113.17, reaching a five-month high. Prices have risen dramatically from below $97 earlier this month due to uncertainties surrounding the resolution of the Iran conflict and its effect on the global energy sector.

In Europe, France's CAC 40 stock index experienced one of the continent's steepest losses, dropping 1.4%. Despite having fewer tech stocks compared to U.S. or Asian indexes, it was negatively impacted by rising yields on French government bonds.

French government bond yields have been climbing due to concerns about the country’s substantial public debt and a strained budget. Ongoing protests in France have heightened pressure on the government to increase its spending, potentially exacerbating its debt levels.

Concerns over high government debts are prevalent across many developed economies, including the United States. Such apprehensions, in conjunction with rising inflation and robust economic signals, have pushed U.S. bond yields to their highest levels in decades recently.

However, on Thursday, U.S. yields fell sharply, indicating that investors were gravitating toward the perceived safety of U.S. government bonds amid the declining tech stock landscape.

The yield on the 10-year treasury note dropped to 4.97%, down from 5.13% the previous day and over 5.30% earlier in the month, marking a significant shift in the bond market.

This decline occurred concurrently with a report suggesting that the U.S. economy may be performing better than expected, indicated by a reduction in unemployment benefit claims last week, signaling that layoffs remain low nationwide.

The health of the U.S. economy has generally supported stronger corporate profits, which contributed to the stock market reaching record highs earlier this month, despite various underlying concerns. On Thursday, two-thirds of the S&P 500 stocks recorded gains, but the substantial declines of influential tech stocks overshadowed them.

Major companies such as Procter & Gamble, Lockheed Martin, and CSX surpassed profit expectations for the summer, helping to soften the market’s overall losses.

Procter & Gamble's CEO, Shailesh Jejurikar, acknowledged ongoing "geopolitical and economic challenges" but maintained the company's growth forecasts for revenue and profits. Consequently, the stock for the brand known for Crest, Olay, and Oral-B rose by 5.6%.

In summary, the S&P 500 declined by 93.74 points, closing at 7,663.43. The Dow Jones increased by 11.89 points, ending at 52,867.98, while the Nasdaq composite dropped 694.00 points, closing at 26,008.86.

Global markets affected as AI stocks plummet and oil prices surge.

A dimly lit global trading floor split in two: on the left, towering screens display AI stock tickers plunging in bright red—Nvidia, Microsoft and Meta logos prominent—arrows slashing downward; on the right, a lone oil barrel and distant rig bathed in warm natural light, a golden price arrow shooting upward; a faint world map etched behind ties both scenes together.

Sharp drops for AI stocks overwhelm markets worldwide as oil prices jump - The Guardian

- Global stock markets fell, led by a 1.2% drop in the S&P 500.  
- AI stocks declined due to borrowing cost concerns and valuation fears despite company growth.  
- Oil prices surged to $113.17 amid uncertainties regarding the war with Iran.

The latest shudder for stocks in the AI industry pulled markets lower worldwide on Thursday, while oil prices hit their highest levels since May on worries about the war with Iran.

The S&P 500 fell 1.2% for its worst day since July, while the Nasdaq composite tumbled 2.6%. The weakness for tech stocks was so overwhelming that it masked strength for the rest of the US market, in which the majority of stocks actually rose. The Dow Jones Industrial Average added 11 points, or less than 0.1%.

AI-related stocks have been swinging on worries that their boom may be under threat because of how expensive it’s become to borrow money to build datacenters and buy more AI chips. Concerns also remain that their prices shot too high following a years-long frenzy, even though companies in the industry continue to report growth.

Lam Research, whose equipment and services help make semiconductors, reported gains in profit and revenue for the summer that were slightly stronger than analysts expected, for example. Usually, that helps boost a stock’s price, but Lam Research’s tumbled 11.8%.

Other big tech stocks likewise fell following quarterly reports that included at least some encouraging nuggets.

Meta Platforms, which is pushing its Muse AI agent, lost 7% after reporting stronger revenue for the latest quarter than analysts expected, shedding tens of billions in market value, though its profit fell short. The company behind Facebook and Instagram also said it was creating a new class of stock to raise money to pay for AI investments and other expenses.

Tesla fell 3.1% after reporting a stronger profit for the summer than analysts expected, though its revenue was weaker than forecast.

The global tide against AI stocks knocked Nvidia down 4.4%. Because the chip giant is the largest company on Wall Street by total value, movements for its stock carry more weight on the S&P 500 than any other, and it was the single strongest force pulling the index lower.

CoreWeave, which has borrowed money to buy Nvidia’s AI chips that it rents out to customers over the cloud, has been particularly shaky in recent weeks because of worries about its debt. It dropped 10.1%.

The weakness for AI and other tech stocks began in Asia, where indexes are full of players in the industry. South Korea’s Kospi tumbled 3.5%, Hong Kong’s Hang Seng dropped 3% and Japan’s Nikkei 225 sank 2.7%.

Also weighing on global stock markets was a jump in oil prices.

The price for a barrel of Brent crude oil leaped 4% to $113.17 after touching its highest level in five months. It’s shot up from less than $97 at the start of the month on uncertainty about when the war with Iran will allow the global energy industry to return to normal.

In Europe, France’s CAC 40 stock index fell to one of the continent’s sharpest losses with a drop of 1.4%. It has fewer tech stocks than Asian or US indexes, but it was hurt by another jump in yields for French government bonds.

Yields have been generally climbing there because of worries about the government’s big debt load and strained budget. Violent protests across France have raised pressure on the government to increase spending, which could add further to its debt.

skip past newsletter promotion Free newsletter | Weekly Sign up to TechScape A weekly dive in to how technology is shaping our lives Enter your email Sign up after newsletter promotion

Worries about big government debt are weighing on almost every developed economy in the world, including the United States. Such concerns, along with high inflation, signals of a still-strong US economy and other factors, have pushed yields in the US bond market recently to their highest levels in decades.

But US yields nevertheless fell sharply on Thursday in a signal that investors were buying US government bonds in search of something safer to own as AI stocks tumbled.

The yield on the 10-year treasury tumbled to 4.97% from 5.13% late on Wednesday and from more than 5.30% earlier this month. That’s a major move for the bond market.

Thursday’s drop in yields came even as another report suggested the US economy may be doing better than economists expected. It said fewer US workers applied for unemployment benefits last week, which could signal that layoffs remain low across the country.

That solid economy has helped US companies broadly to make bigger profits, which is what propelled the US stock market to a record high earlier this month despite the many worries tugging at it. Two out of every three stocks within the S&P 500 rose on Thursday, only to get drowned out by the drops for more influential tech stocks.

Procter & Gamble, Lockheed Martin and CSX were among the latest companies to join the list of those topping profit expectations for the summer, and gains for their stocks helped limit the market’s losses on Thursday.

CEO Shailesh Jejurikar said Procter & Gamble expected “geopolitical and economic challenges” to continue, but it stood by its forecasts for growth in revenue, profit and other measures for the full year. The stock price for the company behind the Crest, Olay and Oral-B brands rose 5.6%.

All told, the S&P 500 fell 93.74 points, to 7,663.43. The Dow Jones Industrial Average added 11.89, to 52,867.98, and the Nasdaq composite tumbled 694.00, to 26,008.86.

Sharp drops for AI stocks overwhelm markets worldwide as oil prices jump

The concerns surrounding job losses due to artificial intelligence (AI) have subsided in recent months after reaching alarming levels earlier this spring. According to Goldman Sachs, the anticipated layoffs might be on hold as the economy braces for a potential recession.

Joseph Briggs, an economist at Goldman Sachs, noted in a recent publication that worries about AI-triggered job displacement have diminished significantly. He projected that while AI could eventually enhance productivity by 15% across the U.S. and other advanced economies, we may see up to 9% of the workforce affected during the transition to AI. This, he explained, combines with a slower pace of hiring that could further impact employment levels by an additional 3%. Essentially, this translates to approximately one in eleven workers potentially facing job displacement.

A crucial aspect of this forecast is the timing. By revisiting historical data on technological impact on employment, Briggs observed that disruptions related to automation tend to be less pronounced when the economy is thriving, but they tend to increase during economic downturns. He emphasized the possibility that AI-related disruptions might exacerbate the cyclical nature of economic activity.

This notion has been reinforced by various analyses from Goldman Sachs over the past year. In October 2025, economists David Mericle and Pierfrancesco Mei suggested that “jobless growth” might become the norm, where economic growth occurs simultaneously with sluggish job creation. They warned that the full ramifications of AI on the labor market might not become evident until a recession is evident.

Briggs corroborated this idea, stating that automation and labor interruptions would likely align with economic cycles. He highlighted a historical trend where periods of economic decline tend to see increased automation as companies look to restructure and reduce labor expenses.

Looking deeper into the matter, Briggs noted that labor market disruptions are more pronounced in unfavorable economic conditions. Automation linked to technological advancements correlates with increases in unemployment rates, with studies showing an increased job loss rate when the unemployment figures are rising.

In his latest assessment, Briggs increased his estimate of the potential job displacement caused by AI from a previously projected 6% to 7% to a more significant 9%, equating to approximately 15 million workers over the next decade. He elaborated that although the narrative surrounding AI employment might sound alarming, the effects would likely be transient, suggesting that any yearly increase in unemployment would remain below one percentage point.

Despite the significant implications of AI on the job market, the urgency surrounding layoffs has quieted for now. Recent metrics indicate that AI-related layoffs have decreased sharply since their peak in the spring, and anecdotal reports from various sectors hint at a subtle shift in hiring patterns. While some industries tied to AI have witnessed a decrease in job growth, brisk hiring in AI-focused infrastructure has somewhat balanced the scales.

Even some of the most vocal critics of AI-related job losses have tempered their rhetoric. Dario Amodei, CEO of Anthropic, who previously claimed that AI may wipe out half of all entry-level jobs, has since acknowledged the complexities introduced by AI. He referred to the Jevons paradox, highlighting how automation could allow the remaining tasks to become more productive, albeit with significant caveats regarding the rapid advancement of AI technologies.

Briggs noted that it's still early in the integration of AI into workplace practices, as only a fraction of U.S. businesses have fully adopted these technologies. Thus far, the overall economy has remained robust, which could have lessened the urgency for companies to make drastic cuts to their labor forces despite projections indicating considerable job losses ahead.

Goldman Sachs warns that AI might replace one in 11 workers, but layoffs are on hold until a recession occurs.

Dimly lit corporate office with rows of empty desks under warm desk lamps; scattered notebooks and idle monitors left behind. On a frosted glass wall looms a large Goldman Sachs logo, above a red downward-sloping line chart. Shadowy business-suited figures drift away toward a stormy skyline visible through a distant window. Overall dark, moody atmosphere signifying jobs lost and a recession on the horizon.

AI could displace one in 11 workers, Goldman says, but the layoffs are likely waiting for a recession - Fortune

- Goldman Sachs predicts up to 9% of workers could be displaced by AI transitions.  
- AI-related job losses may accelerate during economic slowdowns, amplifying overall labor disruption.  
- Current job market impacts appear muted, with fewer AI layoffs than initially feared.  

The AI layoff panic that peaked this spring has gone quiet. Goldman Sachs thinks the layoffs are waiting for a recession.

“Concerns of imminent AI-related labor displacement have cooled in recent months after hitting a fever pitch this spring,” Goldman economist Joseph Briggs wrote in a note published Thursday. Then he laid out what may be coming. Under the bank’s forecast that AI eventually lifts productivity by 15% in the U.S. and other developed economies, he found that up to 9% of workers could be “displaced during the AI transition,” with slower hiring weighing on employment by another 3%. That’s roughly one in 11 workers.

The bigger finding is about timing. Briggs went back through decades of data on how technology has replaced workers and found that “automation-related labor disruption is limited when the overall economy is healthy but accelerates during cyclical slowdowns.”

“AI-related labor disruptions could well amplify the broader business cycle,” he wrote.

Goldman has now made this argument twice, a year apart, from two different teams. In October 2025, Goldman’s U.S. economists David Mericle and Pierfrancesco Mei called “jobless growth” the likely new normal, Fortune reported at the time. “The modest job growth alongside robust GDP growth seen recently is likely to be normal to some degree in the years ahead,” they wrote. Then came their warning: “History also suggests that the full consequences of AI for the labor market might not become apparent until a recession hits.”

Briggs reached the same conclusion. “The timing and extent of automation and labor disruption will likely track the overall economic cycle,” he wrote.

The pattern goes back decades. “Historically automation-related labor disruption is quite cyclical,” Briggs writes, “with companies tending to automate work and shed labor costs when the economy starts to slow.” Mericle and Mei cited a leading explanation last year: “companies use recessions to restructure and streamline their workforce by laying off workers in less productive areas.”

The chart that makes the case tracks the share of Americans in routine jobs: clerks, secretaries, factory machine operators, mechanics. That share “has remained stable during expansions but declined sharply during downturns,” Briggs writes. The line drops in steps at the recessions of the early 1980s, 2001, 2008 and 2020, and it doesn’t climb back to where it was. “Historically most of the labor headwinds from routine job automation are concentrated in economic recessions,” he wrote.

Mericle and Mei pointed to the same pattern last year with the “jobless recovery” that followed the dot-com bust. “During that recovery, despite a soft labor market, productivity growth remained elevated, and GDP growth rebounded earlier than employment growth,” they wrote.

Looking industry by industry, Briggs found that “labor market displacement is more severe when the economic backdrop is less favorable.” The hit to unemployment from a technology-driven productivity gain was 0.1 to 0.2 percentage points bigger when the overall jobless rate was rising. Across 25 occupations that past technology upended, employment fell an extra 4% over five years for every point the national unemployment rate climbed.

“A more concerning takeaway from our analysis is that a cyclical slowdown would likely accelerate AI-related automation and job losses, thereby amplifying the broader business cycle,” he wrote.

How 6% became 9%

Briggs has been raising this number for a while. In August 2025, he and Goldman economist Sarah Dong estimated AI could displace 6% to 7% of the U.S. workforce if widely adopted, within a range of 3% to 14%. That used the same 15% productivity assumption. By June, on Goldman’s Exchanges podcast, Briggs put the figure at about 9%, or 15 million workers over a decade. He first described them as workers who would be “reallocated to new positions.” Thursday’s note calls them displaced.

That’s a long way from where Goldman’s U.S. team stood last fall. Mericle and Mei said then that they were “skeptical of the boldest claims that rapid technological progress could lead to very high unemployment,” citing Goldman’s global economics team, which Briggs co-leads. Briggs can square his estimate with that: he says the displacement is temporary. On the podcast he said that “even under our forecast for nine percent of workers being displaced, we’d still expect that the unemployment rate increase in any given year would be less than one percentage point.”

Part of the answer is in what Briggs counts. Thursday’s note measures gross job losses rather than unemployment, because the jobless rate hides people. “Some workers that are displaced may find jobs relatively quickly,” he writes, “while others may decide to exit the labor force.”

Historically, each 1% gain in technology-driven productivity pushed the unemployment rate up 0.3 to 0.4 points at its peak, fading to zero after two years. Over the same stretch, it raised the cumulative gross job-loss rate by 0.5 to 0.6 points. Multiply the top of that range by 15 and you get 9.

The occupation data shows how bad it can get in one line of work. Briggs looked at 25 jobs hit hard by past technology, including photo processors, movie projectionists, proofreaders, file clerks and bank tellers. On average, employment in those jobs fell 25% in the five years after it started to decline, “suggesting that very large job losses are possible when technology makes labor redundant.” He calls those numbers “likely upper bounds” for AI, since he picked the cases where technology hit hardest.

Most of the shrinkage came from people losing or leaving jobs, more than from employers hiring fewer people. Separation rates rose in 19 of the 25 occupations, by 3.8 points on average. Hiring rates fell in 13, by 0.7 points on average. That’s why Briggs says AI-related layoffs “should provide a good real-time guide” to disruption.

He also lists reasons the 9% could be low. Displacement ran higher during the late-1990s/early-2000s productivity boom, “the period most widely viewed as comparable to today.” It also ran higher after unusually large productivity jumps, and Goldman expects AI’s boost to beat most past ones. “This pattern suggests more scope for displacement during the AI transition,” he wrote.

Why it’s quiet now

“Given that our estimates point to a significant amount of technology-related job losses to come, why has the impact been limited so far?” Briggs asks.

For now, the layoff gauge reads low. AI-attributed layoffs in Challenger’s tracking have pulled back sharply from their spring peak. News coverage of AI job losses has fallen back to levels from the start of 2026. Job growth in the most AI-exposed industries has run about 10,000 below its long-run trend over the past three months, roughly offset by hiring for AI-related construction. “While these patterns suggest that a modest labor market headwind is emerging, the near-term headwinds remain smaller than feared earlier this year,” Briggs wrote.

Some of the most prominent voices warning about AI job losses have softened their tone. Anthropic CEO Dario Amodei spent much of 2025 warning that AI could eliminate half of entry-level white-collar jobs. Onstage with JPMorgan Chase CEO Jamie Dimon at an Anthropic event in New York in May, he reached for the Jevons paradox instead, as Fortune reported at the time. “If you automate 90% of the job, then everyone does the 10% of the job,” Amodei said. “And the 10% kind of expands to be 100% of what people do and kind of 10xs their productivity.” He kept one caveat, and it’s the one Briggs’ note is about: “AI is moving faster than all these previous technologies,” he said. “It’s possible you get these weird behaviors and this big disruption.”

Briggs’ first answer is that it’s still early: only 23% of U.S. establishments have built AI into regular workflows, and most AI-enabled productivity gains remain in the future. His second is the economy. “Given that the overall economy and labor market have remained resilient in 2026,” he wrote, “companies may have had less incentive to focus on aggressively cutting labor costs.”

Generative AI was used for research assistance and/or transcription of this article. The reporter independently reported and verified the factual claims in this article, and a human editor reviewed it before publication.



Google unveils its flagship AI model Gemini 4 following extended delays.

More from AI

Is China primarily utilizing AI tokens for creating micro-dramas?

U.S. Treasury Aims to Curb Transactions Involving Chinese Biotech and AI

Is AI truly better than doctors at predicting cancer survival? I have my reservations | Devi Sridhar

Yen gains strength as AI concerns drive investors toward safe havens.