By Heath Parkes-Hupton
Publication Date: 2026-06-12 02:28:00
One of the world’s leading artificial intelligence companies has highlighted the need for a global framework to halt development, saying systems are getting closer to creating their own successors.
Anthropic, the creator of Claude, has revealed that it is increasingly asking its AI to develop itself – a change from the time when “humans controlled every step in its development cycle”.
The US tech giant, which recently announced plans to become a publicly traded company, said that if the systems were provided “sufficient computing power, this trend points to an AI system capable of designing and developing its own successor completely autonomously.”
“This is called recursive self-improvement. We are not there yet and recursive self-improvement is not inevitable,” wrote CEO Marina Favaro and co-founder Jack Clark in a blog post titled ” When AI builds itself.
“But it could come sooner than most institutions are prepared for.”
In May, Claude wrote 80 percent of the code used by Anthropic – a blast from…