From cancer research to clinical trials, Anthropic’s Claude is making inroads into the field of drug development, assisting biotech companies with tasks ranging from research and rare disease studies to the preparation of regulatory documentation.
Anthropic is expanding the capabilities of its AI assistant, Claude, beyond simple conversation and coding, testing how it can support actual drug development projects.
The company states that its life sciences team is collaborating with six organizations—including Bristol Myers Squibb, BigHat Biosciences, Generate Biomedicines, Divergence Labs, Enveda, and The Column Group—and has also initiated its own internal drug discovery efforts.
Anthropic’s goal is not to replace scientists with a chatbot. Instead, the company aims to determine whether AI can accelerate the complex, multi-stage process of drug development—from identifying potential therapeutic targets and designing molecules to preparing regulatory filings and verifying manufacturing records.
Claude is beginning to assist with drug design.
Drug development involves far more than just discovering an effective molecule. Researchers must continuously refine designs, conduct testing, stabilize production, and demonstrate safety before a treatment can ultimately reach patients.
Anthropic is testing Claude across these various stages, and some early results offer a glimpse into where AI can make a meaningful impact.
At Generate Biomedicines, six scientists used “Claude Code” to propose new designs for genetic instructions used in CAR-T cell therapy—a treatment that enables immune cells to attack cancer.
According to Anthropic, one of these designs was approximately 70% shorter than the standard version and performed better in laboratory tests, helping immune cells produce roughly 10 times more receptors.
At Divergence Labs, Claude contributed to the development of a potential antibody-based therapy for food allergies. The therapy is designed to target and remove IgE—a molecule involved in allergic reactions—from immune cells. Anthropic reports that Claude designed approximately 35,000 different antibodies in just one month.
Researchers then narrowed down the scope of testing to identify which candidates could effectively bind to and remove IgE.
While these are only preliminary results, they illustrate how AI can help researchers explore a much broader range of possibilities before deciding which candidates warrant closer investigation.
AI is also being used to handle paperwork.
Not all bottlenecks in drug development occur in the laboratory. Tasks such as documentation, quality checks, and preparing regulatory filings can consume weeks of a professional’s time. Anthropic stated that at Enveda, Claude reviewed 1,061 pages of production documentation; it not only identified all 11 issues previously found by human reviewers but also uncovered 7 additional ones. The review time was reduced by 89%—from over 25 hours to less than 3 hours.
Claude also drafted regulatory documents—one of which reportedly took only four hours to prepare, compared to approximately three months if done manually.
These results suggest that the contributions of artificial intelligence may extend beyond the discovery of new therapeutic treatments. It can also help researchers spend less time sifting through paperwork, allowing them to devote more energy to evaluating scientific evidence.
From Early Experiments to Viable Medicines
Anthropic is collaborating with Bristol Myers Squibb to develop clinical trial plans. Furthermore, the company has acquired Mithril Medicine, a small firm focused on the research and development of treatments for rare muscular dystrophies.
The project has yielded encouraging results in mice; should the candidate drugs continue to prove effective, human trials could potentially begin as early as the end of 2028.
It is important to note, however, that most of these studies remain in the preclinical stage, meaning the therapies have not yet reached human testing. More sophisticated design and streamlined documentation do not automatically translate into safe and effective medicines.
While Claude may help researchers accelerate their progress, the true test lies in whether these findings can withstand the long and arduous drug development process. In medicine, a promising experiment is merely a beginning, not the end.