Anthropic states that its AI model Claude autonomously identified a novel enzyme system comparable to the machinery powering the gene-editing tool Crispr. This finding represents the initial output from the company’s newly established wet lab and serves as an early assessment of the artificial intelligence’s utility for scientific research ahead of its planned public listing. The company claims Claude located the enzyme system by searching through a vast database of DNA sequences while human scientists provided only the initial prompt and managed the physical laboratory work. Over a period of twenty-one hours, nearly nine hundred fifty Claude agents processed approximately two hundred and ten million tokens before one instance recognised an unusual repeating pattern within the data. The discovery highlights the potential for large language models to perform complex biological analysis without direct human intervention during the search phase.
While the claim of autonomous discovery remains unverified by independent researchers, the result demonstrates a practical application for combining computational power with biological verification. Anthropic aims to use this capability to accelerate drug discovery and genetic engineering processes that traditionally require extensive manual screening. The move also signals a shift towards integrating physical infrastructure directly with generative AI models to validate theoretical predictions in a real-world setting.
- The search involved nearly 950 autonomous agents running for over 21 hours.
- Human scientists were limited to providing the initial prompt and conducting lab work.
- Anthropic is preparing to go public with this new biolab initiative.
