Anthropic wants to develop its own drugs

In this event “Briefing: Artificial Intelligence for Science” Earlier this week, Anthropy Cloud Science announcednew “Artificial intelligence workbench for scientists“Which pulls fragmented tools and data sets into a single environment, generating numbers and visuals. Anthropic, which already dominates the industry with popular coding tools and powerful AI models, framed the launch around what it says is AI’s ability to “dramatically accelerate the pace of scientific discovery and the development of healthcare interventions,” and touted a long list of biotech and pharma clients already using Cloud.

Anthropic also went further, saying it would develop its own drugs. Life Sciences Department Chair Eric Kuderer Abrams He said The company will focus on discovering treatments for “neglected” diseases.

AI companies have been keen to attract science and pharmaceutical clients – OpenAI, Amazon, Googleand others have their own tools and platforms for life sciences. But Anthropic’s planned move is one of the most direct public attempts by a major frontier AI company to develop drugs itself. This puts it in the unusual position of selling software to other potentially competing drug makers. Anthropic joins a broader race that includes AI-first pharmaceutical companies like Insilico, Google DeepMind, Isomorphic Labs, biotech startups, and big pharma that are building or buying their own AI tools.

Anthropic has provided very few specific details about what it hopes to achieve in drug development. At the event, Kuderer-Abrams did not say what the company would do if it found any promising drug candidates. Anthropic did not respond to EdgeRequests for comment seek more details, including which diseases it plans to target first and whether it will collaborate with other companies for lab work, animal testing, clinical trials or manufacturing.

The experts said Edge The uncertainty surrounding Anthropic’s plans reflects broader uncertainty around the AI ​​drug boom itself. “AI drug discovery” can mean many things. “It’s a really broad term,” explained Namshik Han, a professor at the University of Cambridge and co-founder of AI startup CardiaTec. AI is being applied at “every stage of drug discovery,” he said, from finding and improving new compounds to supporting research, data analysis, clinical trials, and even manufacturing. He said every major pharmaceutical company will use AI in some way. Matthew Todd, professor of drug discovery at University College London, echoed the sentiment that AI is already spreading into drug discovery and research, describing it as a “catchy phrase” given its wide range of uses.

There is no doubt that artificial intelligence is changing the development of medicines. Pointing to numerous initiatives by pharmaceutical giants such as AstraZeneca, Novo Nordisk and GSK, Hahn said that AI can already help generate potential drug ideas, such as suggesting new molecules that could interact with parts of the body such as cell receptors that are already known to be involved in a particular disease or are targets for existing drugs. It is extremely useful for accelerating research and helping to “road test” new drug ideas, Todd said. Given Anthropic’s work on pioneering models, the company will supposedly use generative AI to search across vast chemical and biological possibilities and help researchers make connections that might be difficult or slow to find otherwise, potentially suggesting new drug ideas, identifying new disease targets, or finding new uses for existing drugs.

But there is still a long way to go before the AI-designed drug reaches patients. Todd said the field is “a long way away” from getting regulatory approval for an AI-designed drug for human use. He added that the drug discovery process will not take place independently, with human input and supervision needed at all times. Both Todd and Hahn point out that the lack of publicly available high-quality experimental data, such as how different chemicals behave in the body, can slow down drug development efforts as well, emphasizing that even for well-studied areas of biology there are still large gaps in our understanding of how things work.

AI is not in a position to fix many of the slower parts of drug discovery. Frank von Delft, professor of chemical structural biology at the University of Oxford and head of the department of protein crystallography at the Oxford Center for Drug Discovery, said people are right to be excited about advanced AI models, but they are “not yet close to making experiments unnecessary.” Candidate drugs have yet to be tested in the real world to confirm their effectiveness and toxicity and whether they have practical properties that allow them to be safely prepared, stored and delivered as medicines. All of this requires skilled labor, a lot of money, time, and especially clinical work in humans – the point where many promising drug candidates fail. If Anthropic wanted to develop a drug, “it would have to spend a lot on trials,” von Delft said.

Anthropist might be willing to try. Last year, the company It is actively recruiting biologists and building its own wet laboratoriesAs of writing this article, I have several Live applications Hiring for life sciences roles. Anthropic is “actively recruiting” in the region as well, Hahn said, adding that the company has contacted several fellow academics. Without naming names, Hahn said he believes Anthropic has successfully recruited a small number of candidates beyond big pharmaceutical companies and prestigious academic institutions.

With all this complexity, whichever disease Anthropologie chooses, any payoff is likely to be a long way off — at least, the better part of a decade, given how long It usually takes a new drug to go through clinical trials. “There’s always a huge delay” in drug testing, Todd said. “It takes time to experimentally show that something is safe.” No AI-designed drug has ever passed clinical trials and FDA approval to reach the market. Some advanced AI candidates have I entered Clinical trials, but it is difficult to know how much AI will contribute, where it has been used during the process, or whether these candidates outperform traditional drugs. AI can speed up some of the research, but drugs still need to prove themselves the old-fashioned way: through slow, methodical trials conducted in the real world.