Helen Toner says a path to superintelligence is taken for granted in many Bay Area circles and safety cannot rest on one trustworthy leader

The Verge · Interview

Read the source (opens in a new tab)
What happened
Helen Toner, executive director of Georgetown's Center for Security and Emerging Technology and a former OpenAI board member, spoke with Nilay Patel in a Decoder interview published November 14, 2026.
Key points
  • She said the plan to achieve recursive self-improvement "and then we're going to do that until we have superintelligence" is "taken for granted in a lot of circles at this point."
  • She said the most coherent sense of alignment, AI doing what the user wanted, is "not on track to be solved," and called builders' 12-to-24-month deadlines for such problems "terrifying."
  • Rather than a six-month pause, she suggested setting "a higher bar for 'here's how confident we need to be in the safety of these systems before we proceed with something.'"
  • Asked whether Sam Altman is trustworthy, she said "my stance on that is pretty clear" and that if hopes rest on "any one individual being trustworthy, then we're pretty screwed."
Evidence
Published podcast interview with transcript. The statements are her own views and arguments, not policy actions.

See it in the latest record