As concerns over the security implications of increasingly capable AI systems continue to grow, Google DeepMind unveiled the program to support governments, research institutions, and biosecurity organizations in strengthening preparedness against emerging biological threats.
The initiative is being developed in collaboration with Isomorphic Labs, Google’s AI-driven drug discovery company, and focuses on three key areas: pathogen surveillance, accelerating the development of vaccines and novel therapeutics, and improving preparedness for infectious disease outbreaks.
Google says it has established more than 15 collaborations over the past year with governments, biosecurity organizations, and research institutions. It has also expanded controlled access to its advanced AI models for selected researchers, government agencies, and trusted partners. Under Google’s safety framework, these capabilities are released only through carefully managed, limited-access programs rather than being made publicly available.
According to Google executives, the initiative is based on the principle that AI is not only a potential source of biological risk but also one of the most promising technologies for mitigating those risks. As the company develops increasingly powerful AI systems, it is simultaneously investing in safeguards designed to identify and manage potential hazards.
Helen King, Vice President of Responsibility at Google DeepMind, said that if internal evaluations determine an AI model has reached a level of capability that could pose unacceptable biosecurity risks without adequate safeguards, the model will not be released. She added that Google continuously assesses the capabilities of its models throughout development and has not yet reached that threshold.
Company leaders also emphasized that there is broad agreement among leading AI laboratories on the importance of rigorous safety testing before deploying advanced models. They argue that developing increasingly capable AI without robust risk assessment and governance could have serious implications for global biosecurity.
The announcement comes just days after Demis Hassabis, CEO of Google DeepMind and recipient of the 2024 Nobel Prize in Chemistry, called for the creation of an independent international body to establish global standards for advanced AI models. Hassabis has argued that oversight of frontier AI should be systematic and involve industry leaders, technical experts, and government regulators.
Google has invested heavily at the intersection of artificial intelligence and the life sciences in recent years. Its AlphaFold system, which uses AI to predict the three-dimensional structures of proteins, has transformed biological research and drug discovery and played a major role in the work recognized by the 2024 Nobel Prize in Chemistry. Isomorphic Labs is now building on that technology to accelerate the discovery and development of next-generation medicines.
Experts believe AI’s most significant future contributions to healthcare will include rapid pathogen identification, forecasting infectious disease outbreaks, accelerating vaccine development, and enabling more targeted therapies—capabilities that could substantially improve global preparedness for future pandemics.
At the same time, Google stresses that advances in AI must be accompanied by transparent governance frameworks and continuous safety evaluations to minimize the risk of misuse.
Analysts say Google DeepMind’s new strategy reflects a broader shift across the technology industry. Rather than viewing AI solely as a potential biosecurity risk, leading companies are increasingly positioning it as a critical tool for preventing and responding to future biological threats.
Re: ax