The AIxBio Landscape

The report examines how AI accelerates biological research, creating a 'dual-use dilemma' where tools for medical advancement can be repurposed for bioweapons. It analyzes the 'risk chain' of biological weapon development, the capabilities of foundation models and biological design tools (BDTs), and the geopolitical landscape involving the US, China, and Russia. The document proposes a 'preferred' 2030 scenario characterized by geopolitical stability and responsible AI development, offering specific US policy recommendations across biosecurity foundations, technical governance, and international diplomacy.

• AI influences biosecurity risks in two primary ways: by lowering the barriers for non-experts to create biological weapons and by increasing the potential harm (the 'ceiling') of biological agents.
• The 'digitization of biology' has enabled AI to expedite the design of vaccines and diagnostics, but it also allows for the remote development of pathogens via cloud labs and automated biomanufacturing, which can bypass traditional physical security measures.
• Biological Design Tools (BDTs), which are narrow AI models trained specifically on biological data, pose distinct risks from general-purpose foundation models, including the potential to design novel toxins or evade DNA synthesis screening.
• The US government's 'Winning the AI Race: America’s AI Action Plan' (July 2025) shifts focus toward accelerating innovation and dominance, while tasking the Center for AI Standards and Innovation (CAISI) to evaluate national security risks, including biosecurity.
• China is pursuing a 'whole-of-government' approach to AI and biotechnology, with the PLA exploring 'neurocognitive warfare' and the state treating genetic data as a 'national strategic resource'.
• Russia maintains an offensive biological weapons program and has modernized its 48th Central Scientific Research Institute, while using AI for disinformation and aligning S&T initiatives with China and BRICS nations.
• The report's 'preferred' 2030 scenario envisions a world where great power competition is replaced by cooperation, the UN serves as a productive forum for AI governance, and AI-enabled biotechnology is recognized as a public good.
• To achieve resilience, the US should implement mandatory DNA synthesis screening, establish intelligence back channels between AIxBio researchers and agencies, and develop an 'if/then' strategy for BDT capability thresholds that trigger security reviews.
• The report warns of extreme risks associated with 'mirror life' research—lifeforms with reversed chirality—which could potentially circumvent the human immune system and cause irreversible ecosystem disruption.

APA

The Council on Strategic Risks (2025). The AIxBio Landscape. https://councilonstrategicrisks.org/2025/07/31/the-aixbio-landscape/

Chicago

The Council on Strategic Risks. The AIxBio Landscape. 2025. https://councilonstrategicrisks.org/2025/07/31/the-aixbio-landscape/.

Wikipedia

{{cite report |author=The Council on Strategic Risks |title=The AIxBio Landscape |date=31 July 2025 |url=https://councilonstrategicrisks.org/2025/07/31/the-aixbio-landscape/ |access-date=23 October 2026 |via=Climate Insights Directory}}

BibTeX

@techreport{thecouncilonstrategicrisks2025aixbio, author = {{The Council on Strategic Risks}}, title = {{The AIxBio Landscape}}, institution = {The Council on Strategic Risks}, year = {2025}, month = jul, url = {https://councilonstrategicrisks.org/2025/07/31/the-aixbio-landscape/}, urldate = {2026-10-23}, note = {Indexed by Climate Insights Directory} }