During a health-focused event in New York on Tuesday, Google announced that it’s developing a collection of “open” AI models for drug discovery called TxGemma.
The AI models, which Google said will be released through its Health AI Developer Foundations program later this month, can understand both “regular text” and the structures of different “therapeutic entities,” including chemicals, molecules, and proteins, according to the company.
“The development of therapeutic drugs from concept to approved use is a long and expensive process, so we’re working with the wider research community to find new ways to make this development more efficient,” Karen DeSalvo, chief health officer at Google, wrote in a blog post provided to TechCrunch. “[R]esearchers can ask TxGemma questions to help predict important properties of potential new therapies, like how safe or effective they might be.”
Google didn’t say whether the models’ license will allow for commercial use, customization, or fine-tuning. TechCrunch reached out to the company for more information and will update if the company responds.
Countless companies, including Google spin-out Isomorphic Labs, have promised that AI could one day revolutionize drug discovery by dramatically accelerating the earliest R&D steps. While there have been some successes, AI has not provided an immediate magical solution in the lab.
Several firms employing AI for drug discovery, including Exscientia and BenevolentAI, have suffered high-profile clinical trial failures in recent years. Meanwhile, the accuracy of leading AI systems for drug discovery, like Google DeepMind’s AlphaFold 3, tends to vary widely.
Still, big pharma — and investors — appear to be enthusiastic about the tech’s potential. In January, Isomorphic, which has partnerships with pharma giants Eli Lilly and Novartis, said that it expects testing on its AI-designed drugs to begin sometime this year. By one estimate, over 460 AI startups are working on drug discovery, and investors have poured $60 billion into the space so far.