Google DeepMind and Isomorphic Labs unveil joint AI-powered bioresilience program

Google DeepMind and Isomorphic Labs on July 16, 2026 unveiled a joint approach to bioresilience, a program meant both to prevent their AI models from being used to create biological threats and to give scientists, governments and biosecurity specialists tools to defend against future disease outbreaks.

Why the companies are moving into biosecurity

Google DeepMind says the biosecurity landscape is changing quickly, driven by shifts in ecosystems, growing global travel and the risk of AI misuse. At the same time, it calls AI itself a critical tool for responding to these challenges: frontier models and the scientific breakthroughs they make possible should help society become more resilient to future outbreaks.

The work runs on two tracks at once: preventing bad actors from using the models, and providing access to the technology for governments, scientists, biosecurity experts and the companies' own teams.

What has been done over the past year

• Over the past 12 months, the companies have built more than 15 partnerships with government bodies, biosecurity organizations and research groups
• The partnerships aim to prevent model misuse, detect new outbreaks quickly and respond to them promptly
• For models such as Gemini, the companies apply a four-step safety process: threat modeling, evaluations, mitigations and monitoring
• The SynthID watermarking technology is being adapted for biology; it may help DNA synthesis providers filter out potentially dangerous AI-generated biological sequences

Which AI developments underpin the program

The program rests on scientific results the companies describe as breakthroughs: AlphaFold from Google DeepMind, which mapped the three-dimensional structures of nearly all known proteins; Isomorphic Labs' drug design engine IsoDDE (AI-powered Drug Design Engine), which delivers the accuracy needed to work with novel biological systems; and AlphaGenome, which sheds light on the function of the genome.

According to the companies, these systems make it possible to move from simply reacting to natural outbreaks and security risks to proactive defense: faster discovery of therapeutics and quicker, more accurate protection of the global health system. Google DeepMind and Isomorphic Labs provide the models and agents built on them to trusted partners across three areas: preventing, detecting and responding to threats.

To work on risk prevention, the companies bring in in-house biologists and security specialists as well as outside partners, in order to understand potential threats, test the models for resilience against them and build in protective mechanisms.

What this means

Google DeepMind and Isomorphic Labs are effectively formalizing something the AI industry has long warned about: the same models that speed up drug development and understanding of the genome could theoretically be used to cause harm. The public bioresilience program is a major player's attempt to show that the risk is manageable through partnerships with governments and technical measures such as SynthID watermarking, and not only through declarations.

Frequently asked questions

What is the

Google DeepMind and Isomorphic Labs bioresilience program?

It is a joint initiative by the two companies, unveiled on July 16, 2026. It combines measures to prevent the misuse of AI models in biology with tools to help scientists and biosecurity specialists detect and respond to disease outbreaks.

How is SynthID related to biosecurity?

The SynthID watermarking technology, originally built to label AI-generated content, is being adapted for biology. It may help DNA synthesis providers check whether ordered sequences were generated by artificial intelligence and whether they carry risk.