Full Breakdown
Anthropic’s Claude AI Uncovers Potential New Gene-Editing System
By Drooid · · How we work
Core Discovery
On September 23, Anthropic, a U.S. artificial-intelligence start-up, unveiled the first result from its newly created molecular-biology laboratory in San Francisco. Its Claude model identified a previously unknown enzyme system, dubbed “ART,” that resembles the natural bacterial defence mechanism CRISPR. The finding was released as a preprint and has not yet undergone peer review.
How the AI Agents Worked
Anthropic deployed 950 Claude-powered software agents to scan a large DNA-sequence database for 21 hours. The agents catalogued more than 200,000 genes for a single enzyme class and flagged 20 candidates for further laboratory testing—work that Anthropic says would normally require weeks or months of expert effort. Human researchers provided only the initial instructions and later performed the wet-lab experiments; the AI agents autonomously selected the leads to pursue.
Expert Reactions
- “I’m genuinely rooting for all of the frontier labs to seriously get into biological discovery, and I’m excited about what comes out of it,” — Dario Amodei, anthropic chief – Amodei expressed enthusiasm for frontier labs entering biological discovery.
- Feng Zhang, a CRISPR pioneer at MIT and the Broad Institute, described the repeated DNA fragments as “genuinely intriguing” and said they “merit further investigation.”
- Kevin Blake, microbiologist at Washington University School of Medicine, warned that the “CRISPR-like” array does not automatically imply a breakthrough comparable to CRISPR technology. “Because the identified array is ‘CRISPR-like’, some have leaped to conclude Anthropic discovered the ‘next CRISPR’ – ie, a Nobel Prize-winning gene editing technology,” — Kevin Blake “CRISPR-the-technology is very different from CRISPR in nature, which is basically a bacterium’s immune system. It has long been known that there are countless CRISPR-like sequences we have yet to identify and catalogue because there are millions of bacterial species we have yet to study,” — Kevin Blake
Potential Impact and Skepticism
Anthropic suggests the ART system could become programmable like CRISPR, opening avenues for gene-therapy research. The scientific community therefore views the claim as a promising lead that requires extensive validation.
Next Steps
Anthropic noted that a Stanford team has independently described a comparable but distinct system, underscoring the rapid emergence of AI-driven discoveries. The company’s lab continues to conduct human-led experiments on ART, while Amodei indicated that future safeguards could allow AI agents to operate laboratory equipment directly.