While AISI said the Mythos agent had not been instructed specifically to avoid or carry out such behaviour, it was "the first time we have seen risks around autonomy and deception manifest this clearly, without specific prompting, in the real-world".

Anthropic's AI used fake human profiles to trick people in safety test

Generate a cinematic news thumbnail for title:Anthropic's AI used fake human profiles to trick people in safety test. No text in image.

Anthropic's AI used fake human profiles to trick people in safety test - BBC

During a performance at Lil Wayne’s Tha Carter VI Tour on September 29, 2025, Tyga took the stage at State Farm Arena in Atlanta. The San Francisco-based record label Empire has rapidly ascended in the music scene, emerging as one of the largest independent labels in the nation. Their reputation has largely been built upon the success of Shaboozey’s hit "A Bar Song (Tipsy)," which played a pivotal role in the label's acquisition of a historic building in downtown San Francisco. However, Empire is now facing scrutiny due to allegations surrounding the use of generative artificial intelligence in some of its recent music releases.

Fenix Flexin, a prominent Los Angeles rapper and founder of Shoreline Mafia, dropped the track "Rubberz" on June 5 under the Empire banner and Flexxd Up Ent. This piece marks a significant departure from Flexin’s usual minimalistic trap sound, leaning more towards a whimsical Italo disco vibe with vocals that echo the detached delivery of Morrissey. The track has achieved unexpected success, climbing onto the Billboard Top 100, receiving accolades as Song of the Week from the New York Times' “Popcast,” and being included in summer song discussions. Initially, Flexin denied employing AI in his work but later conceded this month through an Instagram comment. The AI tool he allegedly utilized, Treblo, confirmed the creation of the song through its detection technologies.

Tyga's latest album, titled “$TARFACE,” which also comes from Empire, is under similar scrutiny regarding AI involvement. This concept album shifts away from Tyga’s established radio-friendly hip-hop style, channeling an eccentric sound reminiscent of “Rubberz.” Its eclectic mix of ‘80s synth-pop and disco beats incorporates a lo-fi aesthetic, evoking the sounds of retro video game music. Like Flexin, Tyga initially sidestepped discussions regarding AI involvement, boasting about his use of vintage production methods in an interview with People, but later admitted to Vibe that artificial intelligence played a role in shaping his album.

"We definitely used AI as a [tool]. And I know people are going to be mad about me saying this, but it’s where technology is going,” Tyga expressed to Vibe. “It’s no different than when Auto-Tune came out. Some people were opposed to it, but real artists took it and used it.”

In a notable critique, Pitchfork assigned “$TARFACE” a 0.0 rating, marking the first time in two decades that the publication has issued such a score (the last being the notorious review of Jet’s “Shine On”). The review criticized the album as "smug and complacent," describing it as a haphazard blend of ‘80s synth-pop and disco that appears to have been crafted by someone lacking a genuine appreciation for the genres.

Empire has not commented on these recent developments.

San Francisco's Empire record label implicated in AI controversy.

A dimly lit, moody recording studio deep in shadows, its walls lined with cracked vinyl discs spilling glowing, glitchy streams of code; a large “Empire” logo flickers atop a console like a neon sign, while a lone silhouetted executive sits hunched before multiple monitors displaying jagged audio waveforms; a small U.S. flag hangs in the corner, amplifying the tense atmosphere of AI-driven music controversy.

San Francisco record label Empire caught up in AI slop accusations - SFGATE

- Empire Records faces backlash for using generative AI in recent artist releases.  
- Fenix Flexin admitted AI's role in his hit song “Rubberz” despite earlier denials.  
- Tyga's album “$TARFACE” also incorporates AI, receiving a historic 0.0 rating from Pitchfork.  

Tyga performs onstage during Lil Wayne’s Tha Carter VI Tour at State Farm Arena on Sept. 29, 2025, in Atlanta. Julia Beverly/Getty Images

San Francisco record label Empire rocketed to the top of the music industry in recent years, becoming perhaps the biggest independent label in the country. It is best known for releasing Shaboozey’s “A Bar Song (Tipsy),” a colossal success that led the label to the recent purchase of a historic building in downtown San Francisco. However, Empire has now become entangled with accusations of using generative artificial intelligence, as artists behind two recent releases have admitted to incorporating the controversial technology.

Los Angeles rapper and Shoreline Mafia founder Fenix Flexin released the song “Rubberz” on June 5 through Empire and Flexxd Up Ent. It’s a far cry from Flexin’s typical style of minimal trap music, it sounds more like a slice of playful Italo disco, with a crooning, vaguely British vocal that matches Morrissey’s disaffected delivery. The song has become an unlikely hit, charting on the Billboard Top 100, garnering song of the week accolades from the New York Times’ “Popcast” and earning a place on the shortlist for song of the summer. Originally, Fenix Flexin denied the use of AI, but then backtracked and admitted it this month in an Instagram comment. The program he supposedly used, Treblo, confirmed via its detection software that it was generated using Treblo software.

Advertisement Article continues below this ad

Tyga’s new album “$TARFACE,” also released by Empire, faces similar accusations. The concept album eschews Tyga’s typical style of radio-friendly hip-hop for a style that sounds, well, very similar to “Rubberz.” The oddball ’80s synth-pop and disco beats have a distinctly lo-fi element to them, at times reminiscent of vintage video game scores. Like Flexin, Tyga initially dodged discussions of AI, bragging about vintage production techniques to People, before confirming to Vibe that AI indeed had a part in the creation of the album.

“We definitely used AI as a [tool]. And I know people are going to be mad about me saying this, but it’s where technology is going,” he told Vibe. “It’s no different than when Auto-Tune came out. Some people were opposed to it, but real artists took it and used it.”

Pitchfork gave the album a 0.0 rating, the first time the publication has given that rating in 20 years (the last was the infamous review for Jet’s “Shine On”), writing that “$TARFACE is smug and complacent, an amalgam of ’80s synth-pop and disco slapped together by someone who’s seemingly never cared to listen to one synth-pop or disco record in his life.”

BECOME A FRIEND. GET OUR AD-FREE APP. Access our new app with ad-free reading, curated maps and more. GET IT NOW

Empire did not respond to a request for comment.

San Francisco record label Empire caught up in AI slop accusations

Leading dating platforms, including Grindr, are embracing artificial intelligence to enhance matchmaking capabilities, driven by research indicating a decline in overall connections among Americans. NBC News’ Allie Canal discusses how this advanced technology is becoming essential for facilitating relationships on these apps.

Dating apps invest heavily in A.I. to combat 'swipe fatigue' and redefine the pursuit of love.

A warm, sunlit café table with an adult’s hand reaching toward a modern smartphone. On the screen, instead of endless swipe cards, a neat grid of glowing heart-shaped profiles linked by golden, circuit-like lines—AI-driven match suggestions. Soft morning light streams through a nearby window, casting gentle shadows on the wooden surface and illuminating swirling digital hearts emerging from the phone.

Dating apps bet big on A.I. to fight 'swipe fatigue' and reshape the search for love - NBC News

- Major dating apps are integrating A.I. to improve user matchmaking experiences.  
- Studies indicate that Americans are forming fewer connections overall in dating.  
- Tech advancements are becoming central to modern matchmaking processes in dating apps.  

Major dating apps like Grindr are opting for more A.I. in order to make better matches, as studies show Americans are connecting less overall. NBC News' Allie Canal reports on the tech increasingly at the heart of matchmaking on the apps right now.Aug. 17, 2026

Dating apps bet big on A.I. to fight 'swipe fatigue' and reshape the search for love

The ongoing competition in artificial intelligence raises a crucial question: is the race driven by the U.S.'s financial muscle or China’s budget-friendly strategies, largely influenced by state-controlled energy costs? The answer might lie in neither approach.

Greetings from Beijing! I’m Evelyn, and I’m here to share insights from local enterprises in my column, The China Connection.

For China, the battle for AI supremacy is characterized by determination. According to Bruce Liu, CEO of Esoterica Capital, “if there is only one dollar left for China, that dollar will be directed towards AI instead of real estate.” This reflects a broader ambition: achieving self-sufficiency in AI capabilities without dependence on U.S. technology. Liu emphasizes, “It’s not necessary for them to possess the world’s leading AI technology.”

Chinese national policies and regional incentives strategize around this goal. While significant progress has been made in developing advanced chips suited for AI applications, they still do not match the capabilities of Nvidia.

Continuing to solidify its standing, Nvidia has brought together major Wall Street players to back a staggering $500 billion in funding for AI research and development. This highlights a notable capital advantage for the U.S.

A recent analysis by Alexander Kheder, a TMT analyst at BMI, a division of Fitch Solutions, revealed that private sector investment in AI within the U.S. is nearly 23 times greater than that in mainland China. Kheder warns that unless Beijing facilitates access to non-state funding for Chinese AI companies, this funding disparity will likely be a persistent factor in maintaining U.S. leadership.

Despite these challenges, Chinese firms have been launching AI models that deliver comparable functionalities at lower costs, even in light of DeepSeek's recent increase in prices. There is a global eagerness among businesses to explore these options.

However, deploying these AI models still hinges on chip availability—the one area where China currently lags behind the U.S.

Clifford Kurz, a director at S&P Global Ratings, remarked that more financial backing from Beijing could be on the horizon. Yet, he questions the utility of such funding without the requisite chips to support that growth. “There’s nothing to finance if the chips aren’t available,” he stated.

Currently, Huawei's offerings provide only about one-eighth of the computational power that Nvidia can deliver, with most of its resources allocated outside China. Each of Huawei's cutting-edge Ascend 950 chips possesses around 13% of the capabilities found in one Nvidia GB300 chip.

Moreover, Nvidia is set to unveil its even more advanced Vera Rubin chip this year, while Huawei has opted to combine multiple chips to enhance performance. Nevertheless, Kurz estimates that Huawei is projected to manufacture merely 1.35 million advanced AI chips in 2023, which falls significantly short of Nvidia’s conservative output estimate of 6 million chips.

CNBC's The China Connection newsletter: Wealth or influence? The crucial factor in conquering the AI competition

A sleek, sunlit boardroom with a long polished wooden table at its center. On the table sits an oversized chessboard: one side stacked high with gleaming gold coins, the other with a single, glowing neural-network chip pulsating like a futuristic pawn. Behind it, a small Chinese flag on a slender pole catches the light. Floor-to-ceiling windows reveal a modern skyline beyond, and on the wall beside the board drifts a prominent CNBC logo, its reflections dancing across the tabletop.

CNBC's The China Connection newsletter: Money or power? The key to winning the AI race - cnbc.com

- China prioritizes AI investment over real estate to achieve self-sufficiency, says CEO Bruce Liu.  
- U.S. private sector investments in AI vastly exceed China's by a factor of 23.  
- China's chip production lags significantly behind Nvidia, limiting its AI capabilities.

Is the AI race about who can spend the most (the U.S.), or whose price is cheaper (China) thanks in part to state-controlled power costs? It may be neither.

Hi, this is Evelyn, writing to you from Beijing. Welcome to the latest edition of The China Connection — a snapshot of what I'm seeing and hearing from local businesses.

For Beijing, the artificial intelligence race is a story of resolve.

If China has one dollar left, "that dollar is going to be spent on AI rather than real estate," said Bruce Liu, CEO of Esoterica Capital.

The goal is to be self-sufficient in AI, without relying on the U.S., he said. "They don't need to have the best AI in the world."

Everything from China's national policy to district-level subsidies underscores those ambitions. So far, China's made strides in advanced chips for powering AI. But they still fall short of Nvidia's.

Now Nvidia has gathered Wall Street titans to support $500 billion in financing for AI development, demonstrating the U.S. advantage in capital.

Private sector AI investment in the U.S. is around 23 times more than in mainland China, according to Alexander Kheder, TMT analyst, at BMI, a unit of Fitch Solutions.

Unless Beijing makes it easier for Chinese AI firms to tap external, non-state capital, "this financing asymmetry will remain one of the most durable structural explanations for US leadership," Kheder said.

That hasn't stopped Chinese companies from releasing AI models with similar capabilities — at lower prices, even with DeepSeek's weekend price hike. Businesses globally are keen to try them.

But running the models still requires chips — a capability that Beijing lacks compared with the U.S.

China "could announce even more financial support," said Clifford Kurz, director at S&P Global Ratings. "But if they don't have the chips, what's the point of support? There's nothing to finance."

Huawei only offers roughly one-eighth the computing capacity that Nvidia has, mostly outside China, Kurz said. He noted each of Huawei's most advanced Ascend 950 chips has around 13% the computing power of one Nvidia GB300 chip.

Nvidia has an even more powerful Vera Rubin chip coming out this year. And for now, Huawei has compensated by piling more chips together.

But Kurz pointed out the Chinese company is expected to produce just 1.35 million advanced AI chips this year — far less than even the most conservative estimate of 6 million Nvidia chips.

CNBC's The China Connection newsletter: Money or power? The key to winning the AI race

On Monday, ByteDance and the Motion Picture Association (MPA) finalized an agreement aimed at enhancing copyright protections for the Chinese tech company's AI-driven video and image creation tools. This move comes in response to a challenge from the MPA regarding how ByteDance manages intellectual property rights.

The accord specifically addresses ByteDance's Seedance video and Seedream image-generation models, following a cease-and-desist notice issued by the MPA earlier this year. In February, Disney and other industry players expressed concerns that these AI models could produce unauthorized content featuring copyrighted characters and likenesses of celebrities.

In light of these concerns, ByteDance has announced that recent iterations of their models now feature improved mechanisms for protecting intellectual property. These tools are accessible through applications like TikTok, CapCut, and Dreamina.

The MPA's initial concerns were rooted in the potential for Seedance to create unauthorized content featuring well-known film and television characters. Both parties have committed to collaborating on further enhancing protections for copyrighted materials as the landscape of AI technology continues to develop.

ByteDance reaches AI copyright agreement with Hollywood industry association.

A sunlit boardroom overlooking the Hollywood Hills, two suited executives leaning over a sleek glass table to shake hands across a freshly signed contract. Behind them, a large LED screen displays the ByteDance logo in bold, while through the floor-to-ceiling windows you can see the iconic Hollywood sign on the distant hillside. On the table, crisp papers and pens catch the natural light, hinting at a new era of AI copyright cooperation.

ByteDance signs AI copyright pact with Hollywood trade group - NBC News

- ByteDance and MPA signed agreement to enhance copyright protections for AI models.  
- Concerns raised about AI tools generating content with copyrighted characters and likenesses.  
- Future collaboration planned to strengthen safeguards as AI technology continues to evolve.  

ByteDance and the Motion Picture Association on Monday signed an ​agreement to strengthen copyright safeguards ‌on the Chinese company’s AI video and image-generation models, months after the ​Hollywood trade group challenged its ​handling of intellectual property.

The deal, which ⁠covers ByteDance’s Seedance video and ​Seedream image-generation models, follows a cease-and-desist ​letter the MPA sent in February over these tools.

Disney and other studios in February ​raised concerns the AI tools ​could generate content featuring copyrighted characters and celebrity ‌likenesses ⁠without authorization.

ByteDance said newer versions of the models include stronger intellectual-property protections.

Models are offered through services ​including TikTok, ​CapCut ⁠and Dreamina.

The MPA’s challenge came amid concerns that Seedance ​could generate content featuring copyrighted ​film ⁠and television characters without permission.

Both sides said they will continue collaborating ⁠on ​safeguards for copyrighted ​content as AI technology evolves.

ByteDance signs AI copyright pact with Hollywood trade group

For over a year, booksellers have raised concerns that companies in the AI sector are purchasing large quantities of rare books, only to destroy them after scanning for training purposes. This notion has been difficult to substantiate until now. A report by 404 Media reveals that an Airtag discreetly placed in a rare book confirmed the involvement of a major technology company, specifically Amazon, in these bulk acquisitions.

On Monday, 404 Media shared how it partnered with a bookseller to embed an Airtag in a rare book from a significant order. This tracking device subsequently led to an Amazon AI training facility situated in Las Vegas, where a team was reportedly engaged in disassembling books to scan their pages. The facility's branding, featuring a logo of a Tyrannosaurus rex poised to devour a book, underscored the company's apparent disregard for the growing criticism surrounding the practice of destructive book scanning.

In response to these revelations, Amazon chose not to address the specifics of 404 Media’s findings. Instead, the company reiterated its prior position, stating to Ars that it acquires books through commercial means to develop and refine its products and services, without explicitly mentioning AI training.

Nonetheless, Amazon is working on what it terms frontier AI models, which necessitate vast amounts of distinctive training data to maintain a competitive edge against industry leaders like Google, OpenAI, and Anthropic. Currently, many firms are protective of their training datasets to avoid compromising their advantages. The potential to train models using text from rare and hard-to-find books could provide Amazon with a significant benefit, particularly since competitors like Anthropic and xAI have openly declared their avoidance of utilizing rare or antique books for their training processes.

Concealed Airtag exposes Amazon's disposal of rare books for AI training.

A moody, rain-slick back alley at night, an industrial dumpster stamped with a large Amazon logo overflowing with shredded rare leather-bound books and scattered aged pages, a single white AirTag nestled among the wreckage, a warm shaft of amber streetlight illuminating it as a gloved hand reaches in—dark, dramatic investigative feel.

Amazon really is destroying rare books to train AI, hidden Airtag reveals - Ars Technica

- AI firms, particularly Amazon, are suspected of bulk purchasing and destroying rare books.
- An Airtag tracked a rare book to an Amazon facility for destructive scanning.
- Amazon deflects criticism, claiming book purchases are for product development, not AI training.

For the past year or so, booksellers have suspected that AI firms are buying up huge lots of rare books, then destroying them after scanning them to train AI. But this was hard to prove until now, as 404 Media reports that an Airtag hidden in a rare book shows that at least one tech giant, in the race to advance its frontier models, is behind some of the bulk orders: Amazon.

On Monday, 404 Media revealed that it had connected with a bookseller who agreed to plant an Airtag in a rare book that was part of a bulk order. That Airtag was then tracked to an Amazon AI training facility in Las Vegas that housed a team focused on tearing books from their spines and scanning pages, 404 Media reported. Apparently tone-deaf to the escalating backlash over destructive book scanning, a logo on the door of that team’s warehouse, VGT3, showed a Tyrannosaurus rex preparing to devour a book, 404 Media documented.

Amazon deflects

Amazon declined to comment on 404 Media’s findings, only providing Ars with the same statement it gave to 404 Media, which does not mention AI training specifically.

“Amazon purchases books through commercial channels to help develop and improve the products and services our customers use,” Amazon’s statement said.

However, Amazon is developing what it considers frontier AI models, which require a massive amount of unique training data to stay competitive with leading firms like Google, OpenAI, or Anthropic. Right now, firms carefully guard their training data to avoid losing an edge. And training models on text from rare books that are difficult to find would seemingly offer an advantage for Amazon, especially since rivals like Anthropic and xAI have publicly stated that they are not training on rare or antique books.

Hidden Airtag reveals Amazon is trashing rare books to train AI

The political landscape surrounding the artificial intelligence sector is witnessing an unprecedented alliance in Florida, as rival factions that typically vie for influence over AI regulations find common ground through substantial financial backing for a single gubernatorial candidate. This collaborative effort follows their competitive engagements in various primary elections across the United States.

Two prominent organizations, each funded by competing AI billionaires and their associates, have poured more than $2 million into advertisements supporting U.S. Representative Byron Donalds’ campaign to become governor of Florida. Ad analytics firm AdImpact notes that Donalds currently leads both in the upcoming Republican primary and the general election scheduled for November.

This puts Donalds in a pivotal position regarding future AI legislation. The current governor, Ron DeSantis, is somewhat atypical for a Republican as he has advocated for more stringent AI regulations and put forward an "AI Bill of Rights," although this proposal was not passed in the state legislature this year. The next governor of Florida will face the decision to continue DeSantis's regulatory initiatives or lean toward a more lenient national approach favored by former President Donald Trump.

Leading the Future, a super political action committee that promotes minimal AI regulation and enjoys support from influential figures such as venture capitalist Marc Andreessen and OpenAI President Greg Brockman, is aligned with Public First Action, an organization focused on AI safety. This coalition represents a noteworthy shift, as both groups strive to assert influence through support for the same gubernatorial candidate.

Earlier this year, these two groups invested a combined $23 million in a Democratic congressional primary in New York City, which became a battleground over AI governance. Leading the Future alone spent over $8 million to oppose Assemblyman Alex Bores, a proponent of AI safety, while Public First invested more than $15 million to support him. Ultimately, Bores did not win the election.

In the context of the Florida governor's race, their combined spending power is significant, eclipsing that of Donalds’ Republican challengers and dominating advertising efforts across the state.

Bela Sullivan, a research and policy associate at the government ethics organization Issue One, pointed out the escalating competition in campaign funding. “The opposing AI interests are trying to leverage financial power,” she observed. “While they may differ on specific AI policies, they agree that financial contributions allow them access to important decision-makers.”

This strategy mirrors a longstanding practice in Washington, where significant contributions are made to secure influence over upcoming political figures. Aubrey Jewett, a political science expert from the University of Central Florida, reiterated this sentiment: large corporate political action committees seek influence through financial support.

Backed by an early endorsement from Trump, Donalds boasts a strong lead in polling among Republican candidates, including Lieutenant Governor Jay Collins and investor James Fishback. Proponents from both AI committees assert their belief that Donalds, a pro-business conservative representing the Gulf Coast region, is sympathetic to their perspectives, even though his stance remains somewhat ambiguous.

In various discussions, Donalds has expressed his support for a federal AI standard aligned with Leading the Future's interests. Meanwhile, his wife, activist Erika Donalds, echoed sentiments during an appearance on Fox Business, suggesting that bureaucratic interference should be minimized to allow AI businesses to thrive.

Conversely, he has also shown support for DeSantis's regulatory recommendations, emphasizing the need for state measures to safeguard children from potential AI consequences—an argument that aligns with Public First’s goals for localized protective regulations.

Leading the Future has committed over $125 million to assist industry-aligned candidates in the approaching November congressional elections, dedicating $5 million to bolster Donalds’ campaign as early as February. Notably, the advertisements from this super PAC do not reference AI, instead focusing on Trump's endorsement of Donalds.

Meanwhile, Public First has reported raising $80 million to back candidates who advocate for stricter AI control, deploying substantial resources to counteract Leading the Future's influence. This group has begun campaigning for Donalds as of mid-July, highlighting his efforts "for Florida against Washington" concerning the protection of children against AI-related harms.

Chris Stewart, a former Congressman leading Public First’s Republican super PAC, claimed that Donalds “maintains a clear stance on the necessity for states to intervene,” especially given the federal government's slow progress on such matters.

Gates McGavick, Donalds’ communications director, rebutted assertions suggesting the financial influx could sway Donalds towards any specific perspective on AI regulation, insisting that his allegiance lies solely with the electorate of Florida.

While Leading the Future and Public First are focusing their financial strategies on shared candidates in a few smaller races nationwide, there are also areas of contention—such as in South Florida’s Atlantic Coast 25th Congressional District—where the two parties support opposing candidates.

Sullivan from Issue One cautioned that increasing expenditures reflect the urgency behind upcoming electoral decisions that will significantly shape the trajectory of AI policy as public demand for action grows.

This strategic maneuvering aims to establish influence over essential considerations before critical regulatory decisions are made in the rapidly evolving sphere of artificial intelligence.

Rival AI billionaires invest millions in Florida election to influence Trump's future governor.

Two silhouetted, sharply dressed men stand on opposite sides of a giant wooden ballot box carved in the shape of Florida under a stormy twilight sky. Each tips open a metal briefcase etched with subtle circuit-board patterns as stacks of cash spill out in green-tinged streams into the box. A single Florida state flag flutters behind them in a cold breeze, while moody, uplighting casts long shadows and underscores the tense, under-the-surface battle for power.

Feuding AI billionaires pour millions into Florida race to shape Trump’s next governor - Los Angeles Times

- AI industry factions are pooling resources to support Byron Donalds for Florida governor.  
- Donalds is a leading candidate, balancing AI regulation interests from both sides.  
- This spending strategy aims to influence future AI policy amid growing public demand.  

The artificial-intelligence industry’s political machines battling each other over how to regulate the technology are trying a new tactic in Florida after fighting in primary elections across the US: joining together to make multimillion-dollar bets on the same candidate.

Two groups funded by feuding AI billionaires and their allies have each already spent more than $2 million on ads boosting US Representative Byron Donalds’ campaign for Florida governor, according to ad analytics firm AdImpact. The Republican is the leading candidate in both the party primary on Tuesday and the general election in November.

That means Donalds is well-positioned to become an influential voice in the struggle over AI regulation.

Advertisement

Current governor Ron DeSantis is a rare Republican who has pushed for stronger AI regulation and backed an “AI Bill of Rights,” though the measure failed in the state legislature this year. The state’s next governor will have to choose whether to take up DeSantis’s cause or side more closely with President Donald Trump, who favors a light-touch national standard on AI.

Leading the Future, a super political action committee favoring light AI regulation backed by billionaires including venture capitalist Marc Andreessen and OpenAI President Greg Brockman is in the unusual position of backing the same candidate in the governor’s race as Public First Action, an AI safety-focused group backed by Anthropic PBC chief executive officer Dario Amodei and other tech executives.

Earlier this year, the two political machines channeled a combined $23 million into a New York City Democratic congressional primary that became a proxy battle over who will control the future of AI. In that race, Leading the Future spent over $8 million opposing New York Assemblyman and AI safety advocate Alex Bores. Public First’s network spent over $15 million to defend him. Bores lost the race.

Advertisement

In the Florida governors race, though on the same side, they are two of the top spenders, outpacing Donalds’ Republican opponents and dominating the airwaves.

“The two opposing AI interests are trying to out-spend each other,” said Bela Sullivan, a policy and research associate at government ethics group Issue One. “They might disagree on AI policy, but they agree money gets you in the room.”

It’s an old Washington influence tactic: spend early and big to gain access to office-holders.

“This is why big corporate PACs and interest groups give – they want to try to have access and influence with who they think is going to be elected,” said Aubrey Jewett, a University of Central Florida political science professor.

Boosted by an early endorsement from Trump, Donalds has a formidable polling lead in the Republican primary over rival candidates including Lieutenant Governor Jay Collins and investor James Fishback.

Representatives of both AI super PACs claim Donalds, a business-friendly conservative whose congressional district runs along the Gulf Coast of Southwest Florida, is sympathetic to their agenda. But he has maintained some ambiguity.

Advertisement

In private conversations and questionnaires, Donalds has endorsed a light-touch federal AI standard, a key priority for Leading the Future. And his wife, conservative activist Erika Donalds, said in an Aug. 8 appearance on Fox Business that Washington should “get out of the way” so AI companies can quickly expand, echoing an industry talking point.

But he has also privately and publicly supported DeSantis’s approach, arguing states need to step in to protect children from harms linked to AI, an argument advanced by Public First, which advocates for state-by-state AI protections.

Leading the Future, which has pledged to spend over $125 million to help industry-friendly candidates prevail in November’s congressional races, announced plans to spend $5 million on Donalds’ behalf in February. Ads from Leading the Future’s Republican super PAC avoid any mention of AI, instead focusing on Trump’s endorsement of the Republican congressman.

Leading the Future spokesperson Jesse Hunt said Donalds “understands that Florida’s future depends on embracing American innovation.”

Public First, which says it has raised $80 million to support candidates who favor stricter AI oversight, has been spending millions to neutralize or counteract Leading the Future’s influence. Public First Action, a dark-money group backed by Anthropic and other tech executives who favor AI safety guardrails, began spending in support of Donalds in mid-July. The group’s ad lauds Donalds for “standing with Florida and against Washington” in the fight to protect children from AI harms.

Former Utah Congressman Chris Stewart, who leads Public First’s Republican super PAC, said Donalds “has been unequivocal that states should have a role to play there, especially absent the federal government being more urgent in trying to lead on this.”

Advertisement

Gates McGavick, communications director for the Donalds campaign, pushed back on the idea that the wave of financial support could influence the candidate in favor of either side of the AI regulation battle.

“Byron Donalds answers to no one but the people of Florida,” McGavick said.

Leading the Future and Public First are spending on the same candidates in a handful of smaller races across the country, including in Oklahoma and New Jersey. But the two AI super PACs are competing in South Florida’s Atlantic Coast 25th Congressional District, where Public First is supporting former Boca Raton Mayor Scott Singer in his Republican primary while Leading the Future supports his potential general-election opponent, Democratic Representative Jared Moskowitz.

Sullivan, with Issue One, said the spending is escalating because candidates elected this year will likely control the future of how AI is regulated as voters demand action on the emerging technology.

“This is part of a broader effort to shape the political landscape before the most consequential AI-related decisions have been made,” Sullivan said.

Birnbaum writes for Bloomberg.

Feuding AI billionaires pour millions into Florida race to shape Trump’s next governor

Amazon has been acquiring large volumes of books, scanning them to gather data for artificial intelligence training, and subsequently destroying these books. 

An investigation conducted by 404 Media uncovered this operation, which had not been publicly reported before. The team strategically placed a tracking device in a rare book, anticipating that it would be purchased by an AI organization. They then tracked its journey across the country until it reached its final location.

We Followed a Delivery of Rare Books That Concluded at an Amazon AI Training Center.

A dawn-lit shipping dock outside a sleek Amazon AI training facility: an open wooden crate overflows with leather-bound rare books, their gilded spines catching warm natural light. Through large glass panels you glimpse humming server racks and a giant Amazon logo above the entrance. A small drone hovers overhead, scanning barcodes on the volumes—symbolizing the transformation of ancient knowledge into AI data.

We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility - 404 Media

- Amazon is purchasing large volumes of books for AI training purposes.  
- The company scans these books and subsequently destroys them.  
- A 404 Media investigation tracked a rare book to expose this operation.  

Amazon is buying massive quantities of books, scanning them for AI training data, and destroying them in the process.

A 404 Media investigation was able to reveal Amazon’s book buying operation, which hasn’t been previously reported, by placing a tracking device in a rare book we suspected would be acquired by an AI company for training data, and following it around the country to its final destination.

We Tracked a Shipment of Rare Books. It Ended at an Amazon AI Training Facility

Half a year after an AI-generated viral video showcasing a fictional showdown between Tom Cruise and Brad Pitt captivated audiences online, steps have been taken to enhance protections for Hollywood's intellectual property and personal likenesses. 

On Monday, the Motion Picture Association (MPA) announced a new Memorandum of Understanding (MOU) with ByteDance, the Chinese technology leader, aimed at establishing robust safeguards within its platforms, including the AI technology responsible for the viral video.

This agreement specifically targets Seedance, the AI video model behind the Cruise-Pitt clip, along with other products like Seedream 5.0 Lite, TikTok, CapCut, and Dreamina.

The rollout of Seedance and this subsequent arrangement closely echoes a situation involving Sora, a video model from OpenAI that debuted last year. Sora faced similar issues with unauthorized content featuring famous personalities, which ultimately resulted in OpenAI discontinuing the model earlier this year.

In contrast, ByteDance remains committed to further developing Seedance, positioning it as one of the leading video models in the industry, alongside those from Google and others.

Charles Rivkin, Chairman and CEO of the MPA, stated, "This agreement showcases our belief that copyright is fundamental to the film and television industry and reinforces our dedication to safeguarding creative works. We’ve had constructive discussions with ByteDance to put in place meaningful protective measures for Seedance and Seedream, and this MOU signifies our shared resolve to continue collaborating on enhancing these measures."

John Rogovin, the general counsel for ByteDance, commented, "ByteDance acknowledges the intellectual property rights that are essential to creative fields globally. We believe that responsible advancements in AI must coincide with substantial protections for rights holders. We value our productive collaboration with the MPA and its constituents as we strengthen protections across our generative AI offerings. This MOU sets a significant foundation for ongoing cooperation as our technology progresses across various products and platforms."

Motion Picture Association Signs AI Intellectual Property Protection Agreement with ByteDance

A polished wooden conference table bathed in warm natural light; two suited executives shaking hands over a printed contract embossed with an “MPA” seal, its pages subtly patterned with circuit-board traces merging into film-reel icons; a prominent ByteDance logo on a banner behind them; classic film reels and a vintage movie projector resting on a side table, with framed movie posters lining the softly lit background.

Motion Picture Association Inks AI IP Protection Deal With ByteDance - The Hollywood Reporter

- MPA and ByteDance signed an MOU to protect Hollywood's intellectual property and likenesses.  
- The agreement will affect AI video models Seedance and Seedream, among others.  
- ByteDance continues to advance Seedance, emphasizing copyright and responsible AI innovation.  

Six months after a fully AI-generated viral video of Tom Cruise battling Brad Pitt took the internet by storm, the tech behind the ripoff has been tweaked to protect Hollywood intellectual property and likenesses.

The Motion Picture Association on Monday said that it had reached a Memorandum of Understanding (MOU) with the Chinese tech giant ByteDance that will provide “strong guardrails” within its tech and platforms that will protect Hollywood.

The agreement covers Seedance, the AI video model that generated the Cruise vs. Pitt video, as well as Seedream 5.0 Lite, TikTok, CapCut and Dreamina, among others.

The launch of Seedance and the ensuing deal bears a striking similarity to Sora, OpenAI’s video model, which launched last year with a flurry of unauthorized videos featuring well-known characters and people, only to fizzle out after the tech giant agreed to pull back on the ripoffs. OpenAI ended up killing Sora earlier this year.

ByteDance, however, is continuing to invest in Seedance, which has become one of the most advanced video models alongside offerings from Google and others.

“Today’s agreement illustrates our belief that copyright is a cornerstone of the film and television industry – and reinforces our commitment to protect creative content,” said MPA Chairman and CEO Charles Rivkin in a statement. “For the past several months, we have had constructive engagement with ByteDance to implement meaningful guardrails on Seedance and Seedream, and this MOU reflects our shared determination to continue our work together to further fortify those guardrails.”

“ByteDance respects the intellectual property rights that underpin creative industries around the world, and we believe responsible innovation in AI goes hand in hand with meaningful protections for rightsholders,” adds John Rogovin, general counsel of ByteDance. “We appreciate the productive engagement with the MPA and its members as we continue to strengthen safeguards across our generative AI services. This MOU establishes an important framework for continued collaboration as the technology evolves, across a variety of products and platforms.”

Motion Picture Association Inks AI IP Protection Deal With ByteDance

The landscape of artificial intelligence implementation is poised for significant transformation. Traditionally, businesses have been lured into adopting AI technologies thanks to the attractive, heavily subsidized pricing models offered by major enterprise software providers. However, this trend is coming to a close. A notable shift is underway as companies are moving towards models that charge based on actual usage.

Currently, organizations benefit from a unique scenario where they are essentially incentivized to integrate AI solutions. Leading software vendors in the enterprise sector have been absorbing substantial costs associated with essential components like graphics processing units (GPUs), inference capabilities, and tokens. This strategy has fostered a misleading sense of financial and operational stability among executive decision-makers, with various features of premium AI agents marketed as “unmetered,” “complimentary,” or “included.” As this paradigm shifts, companies must prepare for more transparent and usage-based pricing structures in the AI domain.

Strategies for Addressing the Upcoming AI Cost Shock

In a dimly lit office with dark, moody tones, a worried adult executive sits at a cluttered desk bathed only in the warm glow of a monitor showing a red “AI Cost” chart spiking upward. Scattered budget reports, calculators, and a steaming coffee cup lie around, while a large, subtle “AI” monogram looms on the wall behind, framing the scene of financial alarm.

How to Respond to the Coming AI Cost Shock - Harvard Business Review

- AI adoption surged due to subsidized pricing from enterprise software vendors until now.  
- Companies are transitioning to usage-based charges, ending the era of heavy subsidies.  
- Major vendors absorbed costs, creating a false sense of security for executive buyers.  

Today’s AI adoption has been fueled by heavily subsidized pricing from enterprise software vendors, but that era is ending. Companies are shifting toward usage-based charges. As

Today’s AI adoption has been fueled by heavily subsidized pricing from enterprise software vendors, but that era is ending. Companies are shifting toward usage-based charges. As

Today’s AI adoption has been fueled by heavily subsidized pricing from enterprise software vendors, but that era is ending. Companies are shifting toward usage-based charges. As

The corporate AI era is on the verge of a massive shift. Currently, organizations are effectively being paid to adopt AI. Across the enterprise tech landscape, major software suite vendors have been quietly absorbing the steep costs of graphics processing units (GPUs), inference, and tokens. This customer-acquisition playbook created a false sense of budgetary and operational security for executive buyers, characterized by “unmetered,” “complimentary,” or “included” premium AI agent features.

How to Respond to the Coming AI Cost Shock

Sainsbury’s has decided to halt the implementation of AI-driven facial recognition technology at one of its locations after a customer was mistakenly flagged as a shoplifter and asked to leave the premises.

Matt Arnold, a 46-year-old comedy promoter, recounted his upsetting experience while shopping for supplies in East Dulwich, southeast London, ahead of an upcoming standup event. After he completed his transaction using his Nectar card, he was confronted by two store managers who informed him that he couldn't be served due to a prior incident. He was subsequently escorted out, though he insisted on leaving on his own terms to maintain some dignity.

During his exit, Arnold spotted a CCTV monitor displaying a red outline around his face, signifying an alert. He requested the staff hold onto his items so a friend could come by to collect them later for the comedy night. The staff seemed puzzled but complied without recognizing that his behavior was not indicative of a shoplifter, simply adhering to the system’s alerts.

The following day, Sainsbury’s corporate office reached out to apologize and announced a pause on the AI-powered Facewatch system in that specific store while an investigation unfolds. Arnold, however, believes the technology should be suspended in all locations due to concerns over potential misuse and wrongful accusations. He expressed apprehension that vulnerable individuals, particularly those with mental health challenges, could be adversely affected by such incidents.

“There’s a risk that anyone could be falsely accused, especially those who are less able to defend themselves,” he stated. Arnold also criticized the broader use of AI in society, emphasizing a lack of accountability when machines direct human actions without question.

Sainsbury’s employs facial recognition to identify shoplifters and enhance staff safety, but it’s been marred by several controversies. In a similar incident from last September, a customer named Warren Rajah was ejected from a different location after also being wrongly flagged by the Facewatch system. Other retailers, such as Home Bargains and B&M, have faced similar challenges.

Arnold urged Sainsbury’s and other retailers utilizing Facewatch to immediately cease its use until they can ensure its accuracy. While the company has implemented a temporary suspension in East Dulwich, he questions why it wouldn’t also extend to all stores given the centralized nature of the technology and the potential for widespread issues.

Voicing his discomfort with surveillance culture, Arnold acknowledged its necessity in some circumstances but underscored the danger of ceding too much authority to machines, with humans expected to blindly follow their commands.

In response, a Sainsbury’s representative conveyed that the company had reached out to Arnold to apologize for the unfortunate incident, clarifying that it stemmed from human error rather than the facial recognition software, which they assert has a 99.98% accuracy rate. Every match made is reportedly reviewed by a trained manager.

A spokesperson for Facewatch confirmed their technology was not at fault in this instance, affirming that while a correct alert was triggered, the mishandling on-site resulted from human error. They stated that the temporary suspension of the system was a preventive measure to enhance staff training and ensure such occurrences do not repeat.

Sainsbury’s further explained that alerts remain visible on employees’ devices for up to an hour and that all misidentifications are documented for review.

Sainsbury's halts AI scanning following wrongful shoplifting claim.

A dimly lit Sainsbury’s supermarket aisle at night, red laser scanning beams slicing across a stunned adult shopper’s coat, a floating digital exclamation mark signaling a false alarm, a large Sainsbury’s logo on the wall, and a moody “AI SCANNING PAUSED” sign glowing overhead.

‘Humiliated’: Sainsbury’s store pauses AI scanning after false shoplifting accusation - The Guardian

- Sainsbury's paused AI face scanning after a customer was wrongly accused of shoplifting.  
- Customer Matt Arnold felt humiliated and powerless during the erroneous identification process.  
- Sainsbury's attributes the incident to human error, claiming the AI system is highly accurate.  

Sainsbury’s has paused the use of AI face scanning in one of its stores after a customer was wrongly identified as a shoplifter and ejected from the shop.

“I was embarrassed, mortified even, and felt quite humiliated and powerless,” Matt Arnold, 46, said of his ordeal.

The comedy promoter was buying supplies in the store in East Dulwich, in south-east London, for a standup event at Dulwich Hamlet football club when, after scanning his items and a Nectar card, he was approached by two managers who told him he could not be served owing to an earlier incident. He was then asked to leave and they tried to escort him from the store.

“They simply told me they would be walking me out,” Arnold said. “At that point I did stand my ground and said I would be walking myself out. I was determined to keep some dignity. On this point, if nothing else, they did relent.”

As he left, he saw an overhead CCTV monitor alert with a red circle surrounding his face. He asked the shop staff to keep his shopping in the trolley so his friend could come and pick up the supplies for the comedy night happening soon nextdoor.

“I think they were quite confused by this, understandably, but agreed and my colleague Dave went in to pay for and pick up the shop about five minutes later. There was no pause for thought from the staff, no suggestion that they understood this is not how a shoplifter would behave. Just blindly following the machine’s orders.”

Sainsbury’s head office apologised to Arnold the next day and has paused use of its AI-assisted Facewatch technology in the store while an investigation takes place. But Arnold says the technology should be paused in all stores and he is worried about the extent to which AI is being used in society.

“Anyone could be falsely accused and at some point that will be someone vulnerable, someone with mental health issues like anxiety. It’s inevitable,” he said. “Also, I would worry about the confidence-destroying effect of it happening to a younger person or someone less willing or able to stand up for themselves as I have done.”

The supermarket chain uses facial recognition technology to identify shoplifters and other criminals, and says this is to keep staff safe from abuse. But the tech has been plagued with issues. Last September another shopper, Warren Rajah, was ordered out of a branch in Elephant and Castle, south London, after being wrongly identified by the same Facewatch software. Similar cases have occurred including in Home Bargains and B&M stores.

Arnold said: “Sainsbury’s, and indeed any other company using Facewatch, should disable it immediately, at least until they can guarantee the system works flawlessly. They have announced that they are suspending the surveillance in the East Dulwich store, where my incident occurred, but it’s a centralised system, so why not suspend it in every store? If there is a potential problem in East Dulwich – be it with the AI itself or the implementation of it – there are potential problems in all the stores.”

He said he was concerned about the rise of AI in society. “I don’t like surveillance culture, but I can see how at times it is necessary. What isn’t necessary is allowing machines to make decisions and then expect humans to scuttle off and do their bidding unquestioningly.”

A Sainsbury’s spokesperson said: “We have contacted Mr Arnold to apologise for his experience at our Dulwich superstore. The incident was caused by human error, not the facial recognition technology. Customers can be reassured that the Facewatch system has a 99.98% accuracy rate, and every match is reviewed by a trained manager.”

A Facewatch spokesperson said their technology was not at fault in this case. “A correct alert was sent to the retailer, but was subsequently subject to human error in the way it was handled in store,” they said.

The company said suspending a store was a “precautionary measure” to prevent further alerts while staff underwent additional training.

Sainsbury’s said that alerts remained visible on staff devices for up to an hour and it logged all misidentifications.

Anthropic's AI used fake human profiles to trick people in safety test

More from AI

San Francisco's Empire record label implicated in AI controversy.

Dating apps invest heavily in A.I. to combat 'swipe fatigue' and redefine the pursuit of love.

CNBC's The China Connection newsletter: Wealth or influence? The crucial factor in conquering the AI competition

ByteDance reaches AI copyright agreement with Hollywood industry association.