Google Unveils New Open AI Models for Drug Discovery at Health Event

During a healthcare event held in New York on Tuesday, Google announced its initiative to create a collection of «open» AI models for drug development, named TxGemma.

According to the company, these AI models, which are set to be released under its Health AI Developer Foundations program at the end of this month, will be capable of interpreting both «plain text» and various structured «therapeutic entities,» such as chemicals, molecules, and proteins.

«Developing therapeutic drugs from initial concept to approved use is a lengthy and costly process, which is why we are collaborating with the broader research community to explore new ways to enhance the efficiency of this development,» wrote Karen DeSalvo, Google’s Chief Health Officer, in a blog post provided to TechCrunch.

«Researchers can pose questions to TxGemma to assist in predicting key characteristics of potential new treatments, such as their safety or efficacy.»

Google has not clarified whether the models will be licensed for commercial use, customization, or fine-tuning. TechCrunch has reached out to the company for additional information and will provide updates if a response is received.

Numerous companies, including the spin-off Isomorphic Labs, have asserted that AI could one day revolutionize drug development by significantly speeding up the early stages of research and development. However, despite some progress, AI has not yet delivered an immediate magic solution in laboratories.

Several firms utilizing AI for drug discovery, such as Exscientia and BenevolentAI, have faced notable failures in clinical trials in recent years. Meanwhile, the accuracy of leading AI drug discovery systems, like Google’s AlphaFold 3, can vary significantly.

Nevertheless, major pharmaceutical companies and investors are excited about the technology’s potential. In January, Isomorphic, which is collaborating with pharmaceutical giants Eli Lilly and Novartis, announced that it expects to begin trials of its AI-developed drugs this year. According to one estimate, over 460 AI startups are working on drug development, with investors having already poured $60 billion into the field.

Source