Anthropic’s Claude AI Takes Active Role in Its Own Evolution
Published on: September 18, 2026
Anthropic has revealed that its AI model, Claude, is actively participating in developing its successor. This marks a notable step in AI research, where a model assists in creating increasingly sophisticated versions of itself.
According to Anthropic’s statement on September 18, 2026, Claude is responsible for approximately 26 percent of the company’s research and development of its next-generation model. The company emphasized that Claude currently handles many tasks end‑to‑end from high‑level prompts, though still under human supervision.
Beyond that, roughly 90 percent of Anthropic’s R&D work involves Claude working in “collaboration” with humans. In this mode, the model carries out large portions of tasks under close human direction, supporting researchers in iterative design and testing.
This development comes amid growing concerns within the AI community about the pace of advancement. Anthropic CEO Dario Amodei and other leading figures have recently advocated for slowing down AI development to ensure safety and transparency. The company also called for reducing the gap between what leading labs know and what the public understands.
By enabling Claude to assist in designing its next incarnation, Anthropic is exploring a form of recursive AI improvement. While still overseen by human researchers, this could potentially accelerate development cycles. At the same time, it underscores the importance of maintaining robust safeguards as AI systems become more involved in their own evolution.
No comments yet.