Claude now leads a quarter of Anthropic's AI research work
Claude leads 26 per cent of Anthropic's research and can complete most tasks end-to-end under human supervision.
Claude, Anthropic's AI model, is helping the company develop the next, more intelligent version of itself. Claude leads 26 per cent of Anthropic's research and development.
Claude can complete most tasks end-to-end based on broad directions. Humans still watch it work.
About 90 per cent of the company's research happens with Claude's help. Large parts of this stay under close human direction.
The model does not yet work on its own. Anthropic announced this when CEO Dario Amodei and other AI leaders asked for slower development over safety concerns.
The company wants openness on how close labs are getting to recursive self-improvement. This is when a model builds its own successor.
- 26 per cent of R&D
- Claude's research leadership
- 90 per cent
- Research with Claude involvement
- Anthropic CEO Dario Amodei
- Safety concern raised by
Why it mattersAI systems becoming more autonomous in their own development raises important questions about human oversight and if safety measures can keep up.
AustraliaNo direct Australian effect yet, but Anthropic's transparency approach will shape how Australian regulators approach AI governance.
✓ Claims checked against the source and corrected before publish. checked 6 d ago
