is a London-based newsman astatine The Verge covering each things AI and a Senior Tarbell Fellow. Previously, he wrote astir health, subject and tech for Forbes.
Demis Hassabis thinks the world needs an AI watchdog pinch the powerfulness to deed the brakes if frontier models go excessively dangerous.
Writing successful a blog post, the Google DeepMind CEO and cofounder said the US should lead the initiative, arguing that the state is the champion spot to group world standards “given its economical and method standing.” The organization, which could lucifer existing regulators for illustration the Financial Industry Regulatory Authority, would beryllium made up of starring independent experts and representatives from open-source communities and would person the authority to measure frontier models earlier they are released and coordinate an industry-wide slowdown if they were judged excessively risky to deploy.
The blog, titled “A Framework for Frontier AI and the Dawning of a New Age,” based on that the request for world regularisation is becoming much urgent arsenic AI systems turn successful sophistication. Artificial wide intelligence (AGI) “is astir apt only a fewer short years away,” he said. “When we look backmost connected this clip successful the decades to come, I deliberation we will realise we were opinionated successful the foothills of the singularity - thing little than the dawning of a caller property for humanity.”
According to Axios, Hassabis has spent months softly building support for his proposal, including briefing the Trump administration, different AI labs, and European officials, and hopes to person the caller statement up and moving earlier the extremity of the year. He told Axios that “the noises I’ve been proceeding [from the Trump administration] are very positive.”
The connection is the latest effort by Hassabis and different manufacture leaders to found a coherent framework for governing progressively powerful AI systems, arsenic good arsenic mitigate the risks they whitethorn pose. As of yet, location is nary world group of rules governing AI specifically, nor a broad group of rules nationally successful the US. Hassabis, the associated victor of the 2024 Nobel Prize successful Chemistry for his activity connected AI-based macromolecule prediction, besides signed his sanction to a connection calling for tougher protections against AI-aided bioweapons accumulation past month.
Hassabis’ astir caller comments travel a connection from apical economists and tech titans — including Anthropic cofounder Jack Clark and erstwhile Google CEO Eric Schmidt — urging world leaders to return the looming economical impacts of AI seriously.
Follow topics and authors from this communicative to spot much for illustration this successful your personalized homepage provender and to person email updates.
English (US) · 
Indonesian (ID) ·