Deepmind CEO Hassabis says "nobody in the world knows what happens next" so "cautious optimism" means building guardrails now

2026-07-14 · Source: The Decoder · Field: Technology & Digital — Artificial Intelligence & Machine Learning, Emerging Technologies & Innovation · Depth: Novice, quick

Summary

Google Deepmind CEO Demis Hassabis has proposed a detailed framework for governing advanced AI, advocating for a new US standards body modeled after the financial regulator FINRA. This body would develop and implement evaluation protocols for frontier AI models, initially on a voluntary basis, later becoming mandatory, and would be funded by the industry using regularly updated benchmarks. Hassabis, who predicts artificial general intelligence (AGI) is only a few years away with an impact ten times greater and faster than the Industrial Revolution, emphasizes "cautious optimism" given the high stakes and expert disagreement on future developments. His proposal aims to build guardrails now, potentially coordinating development slowdowns if necessary, while exempting non-frontier models to prevent regulatory capture.

Key takeaway

For policy makers considering AI regulation, you should prioritize establishing industry-funded standards bodies for frontier models, similar to FINRA. This approach allows for developing crucial evaluation protocols and guardrails amidst rapid AI advancement and expert disagreement. Focus on creating adaptable frameworks that exempt non-frontier AI to foster innovation while managing systemic risks effectively.

Key insights

Proactive governance frameworks are essential for advanced AI development, balancing rapid progress with inherent uncertainties.

Principles

• Advanced AI development necessitates early, robust governance.
• Industry-funded bodies can establish evaluation protocols for frontier models.
• Differentiated regulation avoids stifling smaller AI innovators.

Method

Establish a US standards body, similar to FINRA, to develop voluntary then mandatory evaluation protocols for frontier AI models, funded by industry and using updated benchmarks.

In practice

• Consider financial regulatory models for AI governance structure.
• Implement a two-tiered regulatory approach for AI models.

Topics

• AI Governance
• Artificial General Intelligence
• Frontier AI Models
• Regulatory Frameworks
• DeepMind
• AI Safety

Best for: CTO, VP of Engineering/Data, Executive, Policy Maker, Director of AI/ML, AI Ethicist

Related on AIssential

• DeepMind CEO again pushes for a frontier AI standards body — A US-led, industry self-regulation body for frontier AI is proposed, sparking debate on its efficacy, international acceptance, and public interest alignment.
• AI Optimism vs. AI Pessimism — AI risk discourse is maturing, focusing on practical economic impacts and governance frameworks over speculative doomsday scenarios.
• Anthropic writes Washington an AI regulation playbook — Frontier AI models necessitate urgent, proactive regulation and societal planning for their strategic risks and economic impacts.
• DeepMind’s Hassabis calls for AI government oversight — Frontier AI development requires external oversight to mitigate escalating risks from competitive pressures.
• Sam Altman goes to the Hill as OpenAI preps policy framework — Rapid AI advancement necessitates urgent policy frameworks, cost optimization, and robust safety measures to manage societal and economic impacts.
• Mythos and the Adolescence of AI Policy — The emergence of powerful AI models like Mythos is rapidly accelerating and complicating global AI policy and risk management.

See Counsel's argued verdicts on the open AI decisions leaders are weighing →

Open in AIssential →

Editorial summary, takeaway, and curation by AIssential. Original article published by The Decoder.