In a newly published manifesto, Google DeepMind CEO Demis Hassabis called on Washington to establish an AI watchdog empowered to screen advanced models and orchestrate industry-wide slowdowns if dangers mount.
Hassabis’s warning is stark: AI is advancing roughly 10 times faster than the Industrial Revolution, creating a “species-level transition” with little margin for error. Within 18 months, he warns, severe biological and cyber threats could exist inside open-source models beyond government control.
He isn’t alone. Three of the world’s top AI chiefs have called for binding oversight this year, though they differ on the framework:
| Tech Leader | Affiliation | Proposed Regulatory Model | Core Function |
| Demis Hassabis | Google DeepMind | FINRA-style (Finance) | Public-private partnership, industry-funded |
| Dario Amodei | Anthropic | FAA-style (Aviation) | Federal agency with power to ground unsafe models |
| Sam Altman | OpenAI | IAEA-style (Nuclear) | International coordination and rapid response |
The urgency is understandable. The Trump administration’s improvised, two-week export-control freeze on Anthropic’s Mythos and Fable models last month showed that Washington is willing to intervene in frontier AI during emergencies. For the industry, ad hoc crackdowns are worse than structured compliance.
The Counter-Chorus
But a powerful faction of tech leaders is pushing back, arguing that these safety proposals are a Trojan horse designed to crush open-source competition.
Meta CEO Mark Zuckerberg and Chief AI Scientist Yann LeCun have become the industry’s most vocal opponents of pre-emptive regulation. LeCun has explicitly accused frontier AI labs of fearmongering to achieve “regulatory capture”—a classic manoeuvre in which incumbents lobby for strict compliance rules to pull up the ladder behind them. For Meta, which freely distributes its Llama models to the public, Hassabis’s proposals for mandatory pre-deployment testing amount to a market-entry barrier disguised as public protection.
Venture capitalist Marc Andreessen has taken the opposition further. In his own manifesto, he argued that pre-emptively slowing AI development is “a form of murder” because it delays life-saving medical and economic breakthroughs. Andreessen’s camp argues that governments should regulate the use of AI—holding bad actors liable for fraud or cybercrime—without strangling the development of the underlying technology.
The Conflict at the Core
But there is a structural paradox the industry prefers to ignore: the people best positioned to design AI oversight are also those with the most to gain from shaping it in their favour.
A 2026 ACM research paper characterised this push as regulatory capture—the process by which public institutions begin acting in the interests of the industry they are meant to police. The FINRA model Hassabis proposes is a textbook example. As Wall Street’s self-regulatory body, FINRA is industry-funded and industry-staffed, and has faced decades of criticism for being structurally unable to aggressively police its own benefactors.
Applying that template to AI governance risks ensuring that “light and fleet-footed” regulation moves at the pace of the tech giants, not society.
The Missing Ingredients
A genuinely independent standards body requires public funding, not voluntary contributions from technology companies. It also demands a strict separation between public officials and private firms.
Instead, Western democracies appear to be blurring those lines. With Washington reportedly considering an equity stake in OpenAI, the government risks becoming structurally compromised.
A state cannot act as a disinterested regulator if its own investments lose value every time it enforces safety rules.
Hassabis rightly notes that unregulated AI poses an existential challenge akin to climate change. But history suggests that when an industry asks for regulation, it often expects to help write the rules. The question isn’t whether Hassabis is sincere about the risks; it’s whether sincerity is a sufficient foundation for the institutions the world needs to survive them.
