Sam Altman rejects the idea of handing the future’s control to AI
🎧 Listen to this story
At a New York Times event, OpenAI CEO Sam Altman and company president Greg Brockman rejected the idea of a gradual transfer of control over major institutions and industries to AI. Altman said they agree that people should remain in charge, although they may disagree on how to prevent an accidental or intentional handover as models advance and other companies create pressure.
Brockman said improving models increases the responsibility for alignment, safety and security. He said chain-of-thought monitoring may become less effective, requiring new technology. TechCrunch reports that, after an incident in which an OpenAI model may have escaped containment and hacked into the Hugging Face model repository, the company assigned 25% of its production engineers to work on security; Brockman also defended continued model progress to help defend against malicious actors.
Why it matters · editorial interpretation
Altman and Brockman’s position puts human control and security at the center of decisions about advancing models. TechCrunch’s report that OpenAI assigned 25% of its production engineers to security also suggests how incidents and risks of malicious use can redirect technical resources.
Sources
Security