Anthropic's AI-Driven Biology Lab Finds a New CRISPR-Like Enzyme System
Anthropic says its Bay Area wet biology lab has found a previously unknown enzyme system in bacteriophage DNA that can cut, copy and paste DNA in ways similar to CRISPR. The company says Claude found it mostly on its own in about 21 hours, running roughly 950 agents that used 210 million tokens, though a similar system had been reported earlier by a Stanford team. Humans did all the physical experiments at biosafety levels BSL-1 and BSL-2, but CEO Dario Amodei has not ruled out Claude running lab equipment autonomously in the future.
Skynet Chance (+0.02%): Large swarms of AI agents making discoveries in biology raise dual-use and bioterrorism concerns, and the stated goal of eventually letting AI run lab equipment itself would give AI more power in the physical world. Keeping humans doing the experiments and limiting the work to low-risk biosafety levels partly offsets this.
Skynet Date (+0 days): The result shows AI agents already doing useful scientific discovery faster than expected, which brings forward the point at which risky autonomous research could happen. At the same time, AI labs are publicly calling for the industry to slow down.
AGI Progress (+0.01%): A possibly new scientific discovery made mostly by coordinated AI agents in 21 hours is evidence of real research and reasoning ability beyond benchmark results. However, the finding builds on earlier Stanford work and still needs outside validation.
AGI Date (+0 days): Using AI to speed up science, and possibly AI research itself, suggests capabilities could grow faster, and Anthropic is clearly putting money into AI-run science labs. The effect is still modest until the discovery is confirmed and repeated.
<< All AI news for September 23, 2026
[ Get the daily index digest on Telegram → ]Related AI News
- Anthropic's Opus 5.5 Beats Larger Models at Lower Cost Amid 'Pacing the Frontier' Pledge 2026-09-22
- Anthropic Confirms It Runs a Wet Biology Lab Where AI Models Direct Physical Experiments 2026-09-18
- Anthropic Names Accenture's Faculty as Its First Embedded Safety Evaluator 2026-09-18
- Amodei Pitches "Pacing the Frontier" Safety Plan as Nvidia's Huang Pushes Back 2026-09-18
- Amodei Floats "Pacing the Frontier" Plan as Labs Debate Self-Policing 2026-09-18