DeepMind CEO Demis Hassabis is calling for an independent standards organization to oversee frontier AI models, modeled after financial industry regulator FINRA. The proposal aims to establish testing protocols and best practices for releasing advanced AI systems.
Hassabis outlined the need for a dedicated regulatory body that would evaluate frontier AI models before deployment. The proposed framework would mirror FINRA's approach to financial industry oversight, creating standardized testing and certification processes.
Under such a system, AI developers would submit frontier models for independent assessment. The standards body would evaluate safety, reliability, and societal impact before companies could release systems to the public.
The proposal addresses growing concerns about AI safety as models become more capable. Recent years have seen increased scrutiny over how companies like OpenAI, Google, and others approach releasing powerful AI tools without clear industry standards.
Hassabis emphasized that an independent body would need adequate funding, technical expertise, and legal authority to be effective. He argued that self-regulation by individual companies has proven insufficient, and government regulation alone moves too slowly for rapidly evolving technology.
The idea aligns with broader industry calls for governance frameworks. Other AI leaders have similarly advocated for regulatory structures, though there remains debate over whether such bodies should be government-led or industry-managed.
A FINRA-style model would involve industry participation while maintaining independence from any single company's interests. This could theoretically satisfy both companies seeking clear guidelines and regulators concerned about safety.
Challenges include defining what constitutes a "frontier" model, determining appropriate testing standards, and ensuring the body has sufficient technical capability to evaluate cutting-edge systems. International coordination would also be necessary to prevent regulatory arbitrage.
DeepMind's parent company, Google, faces its own regulatory pressures globally. The proposal from Hassabis suggests major AI developers recognize the need for formal oversight mechanisms before external pressure forces stricter regulations upon them.
No timeline or specific implementation plan was announced. The proposal remains a conceptual framework rather than an immediate initiative.
Uber's weekly AI agent requests have grown nearly tenfold since February, yet the company has held spending flat since April after exhausting its entire 2026 AI budget in Q1.
A recent paper shows artificial intelligence often diagnoses and treats patients better than human physicians. The findings are prompting difficult conversations within the medical community about the profession's evolving role.
The Relay Q, launching next year, represents the latest push to establish voice as the primary interface for human-computer interaction, challenging the keyboard's decades-long dominance.
An Anthropic researcher demonstrated automated systems that can identify and correct misaligned behaviors without compromising overall performance. The systems improved on all 10 tested benchmarks measuring specific problematic outputs.