The pursuit of artificial general intelligence (AGI) is not merely a competition to be first but centers on controlling the systems that will integrate machine intelligence into everyday life.
This narrative around the AGI race encourages significant investment from major tech companies and their investors, diverting focus from the substantial societal risks involved. If nations continue to view AI as a competitive race, they may overlook the essential challenge of creating societies that can effectively integrate technologies that they cannot fully oversee.
Both public officials and corporate leaders often frame the development of AGI as a competitive sprint—to the extent that the narrative is both comforting and misleading. Rather than a straightforward race toward human-like or superhuman intelligence, the true issue involves how rapidly machine intelligence is permeating businesses, governments, and public sectors, often outpacing the readiness of political frameworks.
Artificial intelligence is increasingly embedding itself in the global economy faster than institutions can adapt, with its benefits likely to consolidate among those who control the computing power and distribution networks. The International Monetary Fund has cautioned that while generative AI may bolster productivity, it could also exacerbate inequality.
In this context, the AGI race metaphor aids leading tech labs and major tech firms in rationalizing massive investments in technology infrastructure. However, this framing often distracts from the profound societal consequences that must be addressed, including job displacement, advanced cyber threats, and threats to sovereignty.
The crux of the matter is not about who achieves AGI first but rather who holds authority over the systems that will make machine intelligence commonplace.
The misconception of the AI 'race'
The prevailing narrative suggests that the U.S. and China are engaged in a fierce competition for technological dominance, with the U.S. holding an edge in advanced labs and semiconductor manufacturing, while China leverages its vast scale and industrial capabilities. This alleged race is often depicted as a matter of mere months separating the two nations from a decisive advantage, but this oversimplifies the complexities governments face.
The definition of AGI is ambiguous, and timelines for its emergence are often unclear. While the U.S. leads in private AI investments, Chinese models are closing the gaps in performance metrics related to cost, speed, and scalability.
The idea that one nation or organization can “win” this race is problematic; different forms of power stem from the most effective models, largest user bases, highest productivity, or most sophisticated surveillance technologies.
Nevertheless, the race narrative supports elevated levels of investment from tech giants, who require a compelling story to justify their expenditures, potentially obscuring the risks faced by broader society as they accelerate toward commercial advancement.
The consequences of AGI expansion will be felt as machine intelligence integrates into governmental, market, and daily life routines, already affecting welfare systems, corporate operations, and military strategies. By the time any formal recognition of an AGI milestone occurs, the landscape of intelligence will have shifted significantly.
Mapping the AI power hierarchy
The race analogy oversimplifies the emerging landscape of AI power. The United States and China occupy the top tier, with differing paths: the U.S. capital-intensive approach contrasts with China’s focus on efficiency and localized production.
U.S. restrictions on chip exports may hinder China but also catalyze its innovations in response to resource limitations. A system geared towards lower costs and broader application could become pivotal in a geopolitical context. For instance, research on DeepSeek’s cost-efficient reasoning model showed competitive outcomes despite the constraints imposed by U.S. GPU export regulations.
Countries in the middle tier are working to establish AI capabilities to reduce dependency on external technologies. Nations like Canada, France, India, Japan, Saudi Arabia, South Korea, Singapore, the UAE, and the UK possess the necessary resources, expertise, or strategic motivations for significant advancements.
Many others will seek limited autonomy in specific sectors like public services, defense, and regulated industries. However, few will construct groundbreaking models, and many may still struggle to maintain control over vital functions due to insufficient computing resources and financial markets.
The lowest tier consists of a majority of nations that will rely on existing models and tools. Open-source frameworks and specialized applications will support government and corporate adaptation, promoting a degree of autonomy, but increasing reliance on systems developed elsewhere.
Interestingly, those willing to leverage open-source frameworks and distributed models may realize the most immediate benefits, as many valuable applications do not necessitate cutting-edge models based in distant data centers. More streamlined systems can significantly improve healthcare diagnostics and fraud detection while functioning closer to operational contexts.
This is critical because general capabilities will emerge through both specialized systems and large models operating in real-world settings. As these systems expand, they can enhance capabilities but also increase vulnerabilities.
The implications for national sovereignty
The current focus on AGI overlooks the interconnected risks that arise, as governments often consider AI and AGI perils in isolation.
National security concerns will shift with the introduction of autonomous weapons and deepfake media. Economic landscapes will be reshaped as technology substitutes for human labor. The very concept of sovereignty will be challenged as nations realize their independence is compromised when the intelligence infrastructure is foreign-owned.
The prevailing investment logic in AGI suggests substantial returns for those holding computing and distribution assets, which may provide solace for investors but raises alarms for workers and taxpayers.
If machine intelligence enhances productivity while diminishing labor income and concentrating wealth, rapid political and social destabilization could ensue.
Few societies are adequately preparing for the scale of disruption on the horizon. Governments continue to adopt narratives centered around innovation and voluntary safety measures, while businesses pursue swift advancements motivated by investor enthusiasm. Citizens are expected to trust institutions that have already begun to lose their credibility.
A labor market upheaval emerges as an immediate political challenge. AI does not have to entirely displace jobs to cause societal turmoil; a reduction in wages and career advancement opportunities in clerical, customer service, and tech-adjacent roles could undermine household incomes and weaken the connections between education, effort, and rewards, which are fundamental to democratic frameworks.
The threat AI poses to democratic structures
Democracies encounter unique challenges since their institutions often falter when it comes to long-term planning. Regulatory frameworks are fragmented, and public trust in experts is dwindling, making it harder for citizens to engage with corporate or political figures.
While China’s centralized system might be more adept at directing investment and mandating innovation adoption, this does not equate to societal resilience or quality of innovation. Instead, it highlights the potential vulnerabilities faced by democratic governance.
The likely trajectory is one of chaos rather than prosperity. The U.S. will persist in pushing technological boundaries to maintain its advantage, while China will focus on disseminating capabilities to circumvent American dominance.
Middle powers will strive for practical independence to reduce their reliance on foreign AI systems. Meanwhile, companies will continue to implement advanced agents quicker than governments can effectively oversee their deployment.
To avoid this chaotic outcome, a more serious discussion around AGI governance is necessary. Leaders need to recognize machine intelligence as a matter of national security and essential state capability rather than merely a component of an innovation agenda.
Strategies for sovereign AI governance should prioritize practical control of critical systems, while labor policies must adapt proactively to displacement. Security governance requires expanding beyond voluntary commitments to enforceable regulations.
The greatest challenge lies in rebuilding collective efficacy within societies that face low levels of trust. Effective governance of AGI will necessitate collaboration among states, corporations, and competing powers, along with public institutions taking action before potential crises arise.
Reassessing perspective on artificial general intelligence
Governments need to reclassify AI as a vital infrastructure rather than merely commercial software, emphasizing the need for robust access to computing resources, public data systems, and evaluation mechanisms. Addressing labor disruption should be framed as both a fiscal and political challenge, with security governance shifting from voluntary agreements to mandatory regulations.
Political leaders must commit to understanding the full implications of general machine intelligence. This involves engaging in meaningful dialogue with technologists, social scientists, and the broader public, as democratic legitimacy relies on involving citizens in discussions about machine intelligence applications.
For most nations, developing a credible sovereign AGI strategy does not entail creating prestigious advanced models but rather ensuring adequate computational resources and trustworthy public data to oversee advanced technologies in critical areas.
Preparation for labor market shifts must accelerate. Governments should proactively identify occupations most vulnerable to disruption and redesign training programs before faced with significant upheaval.
Security governance must adopt a firmer stance, particularly in areas of cybersecurity and defense, necessitating mandatory testing, incident reporting, and enforceable standards—voluntary schemes will be ineffective where rapid innovation is prioritized.
Fiscal policies will likewise require reevaluation. Implementing AI taxes and robot levies could provide funding for societal adaptation. In extreme cases, universal basic income might be essential if employment disruptions become severe, though redistribution alone cannot fulfill the need for work as a fundamental source of dignity and purpose.
