Deepmind CEO Hassabis says "nobody in the world knows what happens next" so "cautious optimism" means building guardrails now
Google DeepMind CEO Demis Hassabis advocates for a new US standards body modeled after FINRA to evaluate frontier AI models. The proposed body could coordinate a slowdown in development if necessary.
- Hassabis proposes a US standards body modeled on FINRA to regulate frontier AI.
- The body could enforce evaluation protocols and coordinate development slowdowns.
- Startups and research models would be exempt from these proposed regulations.
- The proposal aims to balance rapid progress with necessary safety guardrails.
Google DeepMind CEO Demis Hassabis has released a comprehensive proposal outlining a new framework for managing advanced AI systems. Central to this plan is the creation of a United States standards body modeled after the Financial Industry Regulatory Authority, or FINRA. This organization would be responsible for developing rigorous evaluation protocols specifically for frontier AI models.
The proposal suggests that this body would possess the authority to coordinate a slowdown in AI development should safety concerns arise. This measure aims to address the uncertainty surrounding the future trajectory of artificial general intelligence. Interestingly, the plan includes exemptions for startups and smaller research models to avoid stifling innovation outside of major tech labs.
Hassabis emphasizes that while the future of AI is unpredictable, adopting a stance of cautious optimism requires establishing these guardrails immediately. The initiative represents a significant call for structured industry self-regulation backed by government oversight.
Frontier model developers may face new evaluation protocols and potential slowdown mandates.
Companies deploying AI need to prepare for a new regulatory landscape resembling financial oversight.
Regulatory risks and compliance costs for major AI model developers could impact investment strategies.
New oversight mechanisms are being proposed to ensure powerful AI systems remain safe and controlled.
- FINRA
- Financial Industry Regulatory Authority, a non-governmental organization that regulates member brokerage firms and exchange markets.
Vatican hosts Nobel laureates, experts to discuss AI security risks - EWTN News
Australia to establish government AI office to coordinate regulation - Reuters
Major book publishers sue Google over Gemini AI stealing millions of copyrighted works, argue tech giant will ‘weaken the incentive to create’ - amNewYork
BusinessGoogle’s Demis Hassabis says it’s time for a global AI watchdog — led by the US
New York to impose the country's first statewide moratorium on data centers - ABC News - Breaking News, Latest News and Videos
Portugal's AMALIA treats AI as a public good - Peterson Institute for International Economics
Portugal's AMALIA initiative treats AI as a public good, according to the Peterson Institute for International Economics. This approach aims to make AI more accessible and beneficial to the general public.
German media regulator says Google's AI Overviews subject to German media law - Yahoo
German media regulator says Google's AI Overviews subject to German media law Yahoo
Inside China's Governed AI Machine: New Book Explains How Regulation Became AI Infrastructure - PR Newswire
Inside China's Governed AI Machine: New Book Explains How Regulation Became AI Infrastructure PR Newswire
DeepSeek considers new funding round at $71B valuation, barely two months after raising $7B - Crypto Briefing
DeepSeek considers new funding round at $71B valuation, barely two months after raising $7B Crypto Briefing
Saudi AI champion HUMAIN to supply compute to Canada startup Cohere - Semafor
Saudi AI champion HUMAIN to supply compute to Canada startup Cohere Semafor
Grok slurping secrets from private repos, xAI scrambles to fix it - Cybernews
Grok slurping secrets from private repos, xAI scrambles to fix it Cybernews