Researcher, Superalignment Team · OpenAI · 2024
Left the Superalignment team. Said 'I really didn't want to end up working for the Titanic of AI.'
Sworn testimony from the former OpenAI superalignment researcher at the Senate hearing on 'Oversight of AI: Insiders' Perspectives.' Saunders testified that an AGI system could plausibly be built in as little as three years, that no one knows how to ensure AGI systems will be safe and controlled, and that security was not prioritized during his time at OpenAI. His testimony became a touchstone in the debate over whistleblower protections for AI-lab employees.
0 of 1 confirmed
AGI could plausibly be built in as little as three years
“I believe it is plausible that an AGI system could be built in as little as three years.”