Debate grows over whether AI agents can claim scientific discoveries
Anthropic recently announced that its molecular biology lab, powered by 950 Claude agents, identified a previously uncatalogued genetic pattern. However, the claim has sparked intense disagreement among biologists regarding whether pattern recognition constitutes a true scientific breakthrough.
The controversy began after Anthropic reported that its AI agents identified a repeating pattern near a known enzyme after 21 hours of processing. While the company compared the finding to the significance of CRISPR technology, some scientists argue the system merely performed automated laboratory tasks. According to MIT Technology Review, biologist Lucas Harrington noted that identifying gene clusters is often the simpler part of research, whereas determining biological function is the true challenge.
The debate is further complicated by claims of data overlap. The New York Times reported that Mario Rodríguez Mestre of the University of Copenhagen believes his team had already identified this specific pattern. Mestre expressed concern that Anthropic might have learned from his interactions with Claude, a claim Anthropic denies. This tension highlights a broader conflict: whether AI should be viewed as a collaborative tool or an independent discoverer.
Source: When can we say AI made a scientific discovery?
