Anthropic unveiled the first molecular biology result from its new San Francisco laboratory on Wednesday. The company reported that nearly a thousand parallel copies of its Claude AI model screened DNA databases in 21 hours, discovering an unknown enzyme system in bacterial viruses that resembles the CRISPR mechanism.
How Claude Accelerated Basic Research in 21 Hours
The artificial intelligence startup unveiled its molecular biology laboratory result on Wednesday, marking what the company describes as a demonstration that AI can actively accelerate basic scientific research. According to the company’s released details, approximately a thousand copies of the Claude model operated in parallel to sift through a database of DNA sequences over a 21-hour period.
That automated screening achieved an analysis speed that researchers estimate would normally require weeks or months of manual work. Within bacteriophages—viruses that infect bacteria—the AI agents identified a structural pattern that had apparently gone unnoticed until now. The configuration consists of short DNA fragments repeating at regular intervals, positioned directly beside a known enzyme and a protein of unknown function.
This genetic organization mirrors CRISPR, the natural bacterial defense system that scientists previously converted into molecular scissors for gene editing and the foundation of genomic medicines.
Dario Amodei Addresses the UN Security Council on AI Risks
Speaking via videoconference before the United Nations Security Council on Wednesday, the patron of Anthropic Dario Amodei addressed international delegates to plead for international coordination in response to the security challenges posed by advanced artificial intelligence systems.
During his address, the Anthropic leader characterized the newly identified genetic motif with cautious phrasing.
At the same time, Amodei emphasized that such remarkable capabilities carry inherent dangers that require active oversight.
The Mystery of the ART System and Expert Reactions
Anthropic acknowledges that the precise biological role of the newly discovered system, which the team has named ART, remains entirely unknown. In an unpublished preprint that has not yet undergone peer review, the research team notes that experiments have not yet demonstrated whether the associated enzyme is biologically active.
Despite the preliminary nature of the findings, independent geneticists have taken note.
San Francisco Laboratory Protocols and Parallel Research
The biological research team behind the discovery was formed in the spring and operates within the San Francisco region. According to corporate disclosures, the facility does not handle human-infecting pathogens, and human researchers carry out all physical laboratory experiments.
Amodei also revealed that a research group at Stanford University independently described a comparable, yet distinct, system. Looking ahead, the Anthropic patron did not rule out the possibility that Claude models could eventually operate laboratory equipment directly, provided that appropriate safeguards are established.
Broader Biotech Push and the Timeline for Medical Cures
The discovery arrives amid a broader corporate expansion into life sciences. Anthropic launched Claude Science in late June as a specialized workspace environment for researchers and announced independent internal efforts to discover treatments for rare diseases.
The milestone parallels advances from competing artificial intelligence laboratories. Earlier in September, rival firm OpenAI reported a major breakthrough in mathematical research by solving one of the seven Millennium Prize Problems, specifically focusing on fluid mechanics.
Addressing the ultimate horizon of these computational breakthroughs, Amodei reiterated an ambitious objective on social media.
At the same time, the company acknowledges practical limitations, noting that artificial intelligence deployment cannot inherently shorten the required duration of clinical trials.