← Wednesday, November 18 deck

20Research1 source posting

OpenAI Shares New Data on How Its Researchers Use AI, and Calls for RSI Reporting

We're sharing new data on how AI is changing research at OpenAI. We find researchers' AI use continues to rise exponentially, and agents are completing longer tasks, but human involvement remains central in research. Researchers are also spending more time on safety and alignment, with human effort in these areas growing faster than in any other individual research area. This report aims to increase transparency around research acceleration and inform the governance of increasingly capable AI. We believe that AI labs should be required to track and publicly report indicators of recursive self-improvement (RSI): the process by which AI accelerates the development of future AI systems. Our hope is that this expanded report will help sustain and advance shared measurement norms across the industry.

Context

Why it matters

A first party look at how a leading lab's own researchers use models, plus a call for required reporting on recursive self improvement, is a rare official window on research acceleration.

OpenAI posted that it is sharing new data on how AI is changing research at the company. It said researchers' AI use continues to rise exponentially, and that agents are completing longer tasks, but that human involvement remains central in research. It also said researchers are spending more time on safety and alignment, with human effort in those areas growing faster than in any other individual research area, and it argued that AI labs should be required to track and publicly report indicators of recursive self improvement. The figures behind those lines are in the linked company report, not itemized in the post text.

Deep dive

OpenAI's official account said it is sharing new data on how AI is changing research at OpenAI. According to the post, researchers' AI use continues to rise exponentially, and agents are completing longer tasks, but human involvement remains central in research. It also said researchers are spending more time on safety and alignment, with human effort in these areas growing faster than in any other individual research area.

The post frames the report as an effort to increase transparency around research acceleration and to inform the governance of increasingly capable AI. OpenAI said it believes AI labs should be required to track and publicly report indicators of recursive self improvement, which it described as the process by which AI accelerates the development of future AI systems. It said it hopes the expanded report will help sustain and advance shared measurement norms across the industry, and it linked a company page.

Those trend lines and the policy ask are first hand from OpenAI. The post text does not give the underlying percentages, task lengths, or charts. This file does not invent those numbers from the linked page. 'Rise exponentially' and 'human involvement remains central' are OpenAI's own summaries of its internal data, not an outside audit.

Because the lead is official, the existence of the report and the stated findings and policy call are confirmed as OpenAI's announcement. Other labs are not shown here as agreeing to the reporting requirement.

What to expect next

The linked OpenAI report is the place to check the actual charts, definitions, and any caveats on the internal data. Watch whether other labs adopt similar RSI indicators, and whether a regulator picks up the reporting idea. If OpenAI later posts a shorter official summary of the numbers, those figures can be added here with attribution. This should stay a research transparency item, not be folded into the separate safety approach post from the same day.