Artificial intelligence is gradually ceasing to be just a tool for chats and content generation. Anthropic, the company known for its Claude model, has announced the launch of the Claude Science scientific platform and unveiled ambitious plans to independently develop drugs. This concerns medications for so-called "neglected" diseases — those often overlooked by pharmaceutical giants due to low profitability.
The Claude Science platform was presented at the event The Briefing: AI for Science. Based on the description, this is not just another AI assistant for scientists, but a full-fledged environment that combines disparate tools, datasets, and visualization capabilities. Anthropic has already stated that the product is being used by clients in the biotechnology and pharmaceutical industries.
From Software Provider to Drug Developer
The key difference between Anthropic and its competitors — OpenAI, Amazon, and Google — is that the company is not limiting itself to the role of a software provider for medicine. It has publicly stated for the first time its intention to independently develop drugs. However, when directly asked about specific research areas and types of diseases, the company gave evasive answers.
Nevertheless, indirect evidence points to the seriousness of these intentions. Over the past year, Anthropic has been actively hiring biologists and building its own laboratories. As Cambridge University professor and CardiaTec co-founder Namshik Han said, the company approached his colleagues in academia with recruitment offers. Moreover, the AI project has already poached some specialists from giants such as AstraZeneca, Novo Nordisk, and GSK.
AI in Medicine: Acceleration, Not Replacement
Experts interviewed during the event confirm that artificial intelligence is already being actively used at various stages of drug development — from searching for and optimizing chemical compounds to analyzing clinical trial data. However, as University College London professor Matthew Todd notes, "AI-driven drug development" is too broad a formulation. In practice, such systems rather accelerate research and help pre-validate hypotheses, but they do not eliminate the need for experiments and human labor.
Oxford University professor Frank von Delft emphasizes that even with AI, potential drugs still need to be tested for efficacy, toxicity, and storage suitability. And the most expensive stage remains clinical trials. Scientists agree that it is still a long way off before "AI drugs" approved by regulatory bodies appear — due to the complexity of obtaining permits and the lack of high-quality open data.
Expert opinion. Anthropic's move is a logical development for a company that wants not just to sell "shovels" to gold prospectors, but to mine gold itself. However, pharmaceuticals are not IT: here, the development cycle is measured in years, and the cost of a mistake is patients' health. If Anthropic manages to combine the speed of AI with the stringent requirements of regulators, it could change the rules of the game. But for now, this is more of a bold statement than a realized strategy.