YouCongress
About Sign up

Comment by Demis Hassabis

Avatar
Nobel laureate, AI Researcher and CEO of DeepMind
The risk from the technology itself, as it becomes more autonomous, more agent-based, is what's going to happen over the next few years. We have to make sure we keep control of the systems, that they're aligned with our values, that they're doing what we want, that it benefits society. AI Unverifiable source (2025)
Like Share on X 2mo ago
Policy proposals and claims
Require human-in-the-loop oversight for agentic AI systems acting in high-stakes domains

Verification History

AI Unverifiable Year 2025 is acceptable. Quote attributes to Demis Hassabis (DeepMind CEO) concern about increasingly autonomous, agent-based AI and the need to keep control of systems so they're aligned with our values and benefit society. The vote "for" on statement 418 (Require human-in-the-loop oversight for agentic AI systems acting in high-stakes domains) aligns correctly with this stance. Independent web search corroborated that this is from Hassabis's 60 Minutes interview where he discussed AI becoming more autonomous/agentic, needing guardrails, control, and alignment with our values. However, the source_url (cbsnews.com 60 Minutes transcript) returns HTTP 403 Forbidden and blocks WebFetch, so I could not directly confirm the exact quote text on the page. Marking ai_unverifiable per protocol for blocked sources, though external corroboration is strong. · Hector Perez Arenas claude-opus-4-8 · 8d ago
replying to Demis Hassabis