Google DeepMind Launches Bioresilience Program to Counter Biothreats with AI

Google DeepMind has announced a new bioresilience program, developed in partnership with Isomorphic Labs. The initiative is designed to assist governments and researchers globally in the prevention, detection, and response to biological threats Source.

This move comes shortly after DeepMind CEO Demis Hassabis advocated for the establishment of a standards body for frontier AI, emphasizing the company's commitment to developing AI guardrails.

AI for Global Biosecurity

The bioresilience program aims to improve pathogen surveillance, accelerate the design of vaccines and therapeutics, and strengthen responses to outbreaks. Google DeepMind has also clarified that it considers future biological applications of AI as 'critical capability level' features requiring specific mitigations before public release.

Strategic Partnerships and Access

Google DeepMind has established over 15 partnerships in the past year with various governments, biosecurity organizations, and research bodies. As part of this program, Google plans to expand access to its AI models and agents for trusted researchers, governments, and biosecurity partners. These releases are classified as "low-risk" restricted deployments under Google's safety framework, not public ones Source.

Helen King, Google DeepMind's Vice President of Responsibility, stated that if a critical capability level were reached without appropriate mitigations, products would not be launched. She noted that this threshold has not yet been met, citing ongoing evaluations during the development process.

Industry Alignment on Safety

There is a consensus among frontier AI labs regarding the necessity of rigorous testing before releasing models, especially when dealing with biological capabilities. Owen Larter, Google DeepMind's Senior Director of Frontier AI Policy and Global Affairs, expressed optimism that the Trump administration's executive order process will lead to a robust framework for testing frontier AI models before their deployment.

Google's History in Biosciences

Google has a track record in AI applications within biology. Its AlphaFold model, which predicts protein structures, has been instrumental in disease understanding and drug development. This work earned Demis Hassabis a share of the 2024 Nobel Prize in Chemistry Source. Isomorphic Labs, a Google affiliate, is now focused on commercializing this AI expertise for drug discovery.

Broader Regulatory Context

This biosecurity effort aligns with Hassabis's recently proposed governance framework. Hassabis previously suggested a more "systematic" approach to AI regulation, funded by the industry, staffed by technical experts, and accountable to the U.S. government. This underscores Google's evolving perspective: rather than viewing biology solely as an AI safety concern, the company is increasingly positioning frontier AI as a vital tool for biosecurity.