OpenAI has officially entered the field of AI-assisted drug discovery, challenging competitors such as Google and triggering sharp declines in the stock prices of related publicly traded companies. OpenAI announced the launch of GPT-Rosalind, an artificial intelligence model specifically designed for life science research, aimed at accelerating the drug discovery process and assisting researchers in handling large volumes of data and translating scientific achievements into clinical applications. The model is currently available as a research preview to select enterprise customers, with initial users including pharmaceutical company Amgen (AMGN), vaccine manufacturer Moderna (MRNA), and the nonprofit bioscience organization Allen Institute. Following the announcement, the share prices of multiple publicly traded companies involved in drug development fell sharply.
According to Joy Jiao, head of life sciences research at OpenAI, GPT-Rosalind is designed to serve as a research collaboration tool for enterprise users, particularly for biology research that increasingly relies on computational processing. In a media briefing, she stated that AI can genuinely help researchers accelerate the most complex and time-consuming steps in scientific workflows. However, Jiao also emphasized that OpenAI does not currently believe AI can independently propose new disease treatment regimens. This statement, to some extent, delineates the application boundaries of the model—assisting acceleration, not leading innovation.
OpenAI’s move is the latest action in the accelerating layout of AI applications in healthcare and science by technology companies. OpenAI, Anthropic, and Alphabet’s Google have in recent years continuously increased their investments in AI applications in science and healthcare, covering areas such as assisting new drug research and analyzing personal medical data. In 2024, two scientists from Google DeepMind jointly won the Nobel Prize in Chemistry for AlphaFold, an AI system capable of predicting protein structures, further raising industry attention to the potential of AI in life sciences. Currently, some drugs discovered with the help of AI have entered early-stage clinical trials, but overall, the technology in this field remains in its early stages of development.
As AI capabilities continue to improve, concerns over potential misuse have also grown. Yunyun Wang, head of life science products at OpenAI, stated that in addition to evaluating whether institutions can use the new model safely, the company has built into the system a “high-precision flagging” mechanism that automatically issues alerts when users approach specific indicators or thresholds related to biological weapons. OpenAI and Anthropic are currently competing to develop advanced AI models capable of undertaking broader tasks, spanning programming, scientific research, and cybersecurity, in hopes of convincing more enterprises to adopt the technology to save time and costs. However, as technical capabilities advance, particularly breakthroughs in software development, concerns over AI misuse have also risen.
Additionally, OpenAI has released an update to its Codex product. The announcement states that Codex can now operate your computer collaboratively, work with more tools and applications used in daily life, generate images, remember preferences, learn from past actions, and handle continuous and repetitive tasks. This update will be gradually rolled out to users of the Codex desktop application logged in via ChatGPT. For enterprises and developers, this direction points toward higher degrees of automated workflows and stronger toolchain integration capabilities.
