Updated
Updated · MIT Technology Review · Sep 28
Biologist Says Anthropic's 950-Agent Biology Finding Was Already Known
Updated
Updated · MIT Technology Review · Sep 28

Biologist Says Anthropic's 950-Agent Biology Finding Was Already Known

3 articles · Updated · MIT Technology Review · Sep 28

Summary

  • Mario Rodríguez Mestre said his University of Copenhagen team had already identified the DNA pattern Anthropic recently presented as its AI lab’s first molecular-biology “discovery.”
  • Anthropic’s 950 Claude agents had spent 21 hours sifting data and flagged a repeating pattern around a known enzyme, not a new sequence or a demonstrated biological mechanism.
  • The claim drew criticism from biologists, including Lucas Harrington, who argued that spotting unusual gene clusters is often preliminary work and not the hard part of establishing a true discovery.
  • Mestre also questioned whether Anthropic learned from his prior chats with Claude; Anthropic denied that, but he said he is stopping use of the chatbot.
  • The dispute highlights broader skepticism toward AI companies framing tool-assisted scientific screening as autonomous discovery, a standard critics say risks overstating progress and blurring how science advances.

Insights

If an AI finds a DNA pattern but a human must prove its function, who really made the discovery?
Will the race to crown AI as independent scientists fundamentally redefine what counts as a scientific breakthrough?
How can AI companies prove their models made independent breakthroughs without secretly copying data from users?