OpenAI launches life science AI model GPT-Rosalind, challenging Google DeepMind's dominant position in drug discovery.

OpenAI is launching an early version of an artificial intelligence model designed to accelerate drug discovery, joining a hot area of focus for current tech companies - all of which are eager to prove that AI can pave the way for more scientific breakthroughs. ChatGPT developer OpenAI announced on Thursday that its new model, GPT-Rosalind, is specifically designed for life science research, such as helping extract insights from Beijing Vastdata Technology and translating scientific research into patient-centered healthcare applications. The model will initially be provided in a preview version to some enterprise clients, with the first users being pharmaceutical companies like Amgen, vaccine manufacturer Moderna, and the nonprofit biomedical research organization, the Allen Institute. OpenAI, Anthropic PBC, and companies under Alphabet such as Alphabet Inc. Class C are increasing investments in the application of artificial intelligence in the fields of science and healthcare, ranging from guiding new drug development to analyzing individual medical data. In 2024, two scientists from Alphabet Inc. Class C's DeepMind received a Nobel Prize in Chemistry for AlphaFold - an AI system that predicts protein structures. This technology is currently seen as promising but in its early stages, with some AI-developed drugs already in early clinical trials. The deep learning model GPT-Rosalind customized for life science research released by OpenAI is seen as a direct challenge to Alphabet Inc. Class C's DeepMind dominant position in the field of biological sciences. Named after the biophysicist Rosalind Franklin, the model represents a significant breakthrough for OpenAI in complex scientific reasoning abilities, moving from simple text generation to deep involvement in evidence synthesis, hypothesis generation, and multi-step experimental planning in scientific research, completely breaking the previous "illusion" limitations of generative AI in the rigorous scientific field. To quickly build an industry moat, OpenAI also disclosed a series of heavyweight industry alliances. In addition to a joint research and development partnership with Eli Lilly for multidrug-resistant bacteria, OpenAI has received significant support from pharmaceutical giant Amgen and vaccine research pioneer Moderna this time. These partners will use the powerful biochemical analysis capabilities of GPT-Rosalind through a "controlled access" model to compress the time and cost of early drug discovery. Joy Jiao, head of life science research at OpenAI, expressed the company's hope that the model will become a "research partner" for enterprises in the field of biological research, especially in computer-dependent biological work. Although OpenAI does not currently believe that AI can independently propose new disease therapies, she stated in a press conference, "We do believe that AI can help researchers complete the most complex and time-consuming parts of the scientific process more quickly." It is worth noting that OpenAI is showing greater ambitions in its business model, with CEO Sam Altman hinting that the company will no longer settle for subscription fees in the future, but rather lean towards sharing research and development risks with pharmaceutical companies and extracting technology royalties once new drugs are approved, deeply involving in the distribution of the trillions of dollars in annual pharmaceutical value. In response to the news of the OpenAI model, companies engaged in drug discovery saw a significant drop in their stock prices. IQVIA Holdings Inc. fell 3.45% on Thursday, Charles River Laboratories International, Inc. dropped 2%, Recursion Pharmaceuticals fell 4%, and Schrodinger Inc. fell nearly 3%. These dynamics indicate that the AI drug development industry has entered a highly competitive phase. For a long time, Alphabet Inc. Class C's DeepMind has been seen as the undisputed benchmark in the field of protein structure prediction with its AlphaFold series models. However, OpenAI is attempting to provide a more flexible and proactive research assistance solution by integrating large-scale language models with specialized biochemical knowledge, aiming to challenge AlphaFold's dominance.