OpenAI launched the Rosalind Biodefense program to give selected developers and government partners access to the GPT-Rosalind life sciences AI model, designed to reason about molecules, proteins, genes, and disease biology to accelerate hypothesis-to-experiment workflows.

The program is briefing the White House and federal agencies, with public-health-focused agencies actively involved in shaping the approach.

Access to GPT-Rosalind is tightly controlled and selective, starting with public health-focused agencies and extending to allied international partners.

The initiative builds on existing safety measures and targets projects that use AI to accelerate defensive research, including literature synthesis, protocol design, model-building, data harmonization, simulation, and decision support.

Applications are open to projects with clear public-benefit goals in biodefense or pandemic preparedness, focusing on the listed capabilities.

Related advances include a February release noting GPT-5-driven autonomous lab capabilities that reduce cell-free protein-synthesis costs by about 40% in a full-system setup with automation and human input.

There are two tracks: a developer track for epidemiological modeling, early detection, screening, and non-pharmaceutical interventions; and a government track for workflows like early-warning systems, outbreak-response planning, diagnostics, and medical countermeasure development.

The initiative acknowledges AI’s biosecurity implications, balancing potential defense benefits with concerns about enabling biological threats.

The overarching objective is to strengthen biodefense and pandemic preparedness by enabling early warning, diagnostics, and vaccine development while addressingAI-driven bioweapons concerns.

Early partners include Lawrence Livermore National Laboratory, Johns Hopkins Applied Physics Laboratory, and CEPI, with Fourth Eon and SecureDNA applying the model to DNA screening.

Key collaborations include LLNL pairing GPT-Rosalind with supercomputing for medical countermeasures, JHU APL integrating it into protein-engineering platforms, and Coalition for Epidemic Preparedness Innovations using it for the 100 Days Mission and faster vaccine development, including work on Bundibugyo Ebola in DRC and Uganda.

The funding and policy context features Biden-era budget discussions on defense and AI investments, with biodefense funding facing trims in some areas, alongside OpenAI’s Preparedness Framework classifying high-capability biology AI use and safeguards.