OpenAI says it's making 'strong progress' toward an automated AI researcher, admits it doesn't yet know how to safely reach aligned full RSI

ZeroStateReflex · x · 2026-09-07

Key takeaways from OpenAI's latest announcements: the company says it is making strong progress toward an automated AI researcher, and chief scientist Jakub Pachocki expects the current pace of progress to extend into RSI (recursive self-improvement). The automated researcher will work under human supervision, with a north star of keeping humans in the self-improvement loop; OpenAI notes it could double as an automated safety/alignment researcher. Decisions on rapid RSI depend on preserving human control and informed democratic choices — and OpenAI admits it does not yet know how to safely reach aligned, full RSI.

Related event: OpenAI Says It Hit Automated Research Intern Milestone, Eyes Fully Automated Researcher by March 2028(11 posts)→

Original post →

More from AGI Musings

AGI Musings channel →