Models that improve on their own are AI's next big thing

"Known technically as recursive self-improvement, the approach is seen as a key technique that can keep the rapid progress in AI going. Google is actively exploring whether models can "continue to learn out in the wild after you finish training them," DeepMind CEO Demis Hassabis told Axios during an on-stage interview at Axios House Davos. Sam Altman said in a livestream last year that OpenAI is building a "true automated AI researcher" by March 2028."

"A new report from Georgetown's Center for Security and Emerging Technology shared exclusively with Axios shows how AI systems can both accelerate progress while making risks harder to detect and control. "For decades, scientists have speculated about the possibility of machines that can improve themselves," per the report. "AI systems are increasingly integral parts of the research pipeline at leading AI companies," CSET researchers note, a sign that fully automated AI research and development (R&D) is on the way."

"The idea of models that can learn on their own is a return of sorts for Hassabis, whose AlphaZero models used this approach to learn games like chess and Go in 2017. Yes, but: Navigating a chessboard is a lot easier than navigating the real world. In chess, it's relatively easy to logically double check whether a planned set of moves is legal and to avoid unintended side effects. "The real world is way messier, way more complicated than the game," Hassabis said."

Recursive self-improvement enables AI models to iteratively improve after initial training, sustaining rapid progress. Google is exploring whether models can continue to learn in the wild after training, and OpenAI aims to build an automated AI researcher by March 2028. Automation in AI R&D can accelerate innovation while making risks harder to detect and control. AI tools are becoming integral parts of research pipelines. Current policy visibility into AI R&D automation is limited and often relies on voluntary disclosures. Recommended responses include increased transparency, targeted reporting, and updated safety frameworks, while poorly designed mandates could backfire.

Read at Axios