An interactive map of the people, institutions, funding, and research connecting biosafety, biosecurity, and AI.
AI is transforming biology at an extraordinary pace — from protein design to genome editing to pandemic forecasting. That creates both tremendous opportunity and real risk. Yet the landscape of who is working on these problems, who is funding them, and how the research community is connected has been surprisingly hard to see in one place.
The Biosecurity Atlas makes that landscape visible. It connects researchers from major AI-biology workshops to the grants, programs, and institutions behind their work, spanning government agencies (DARPA, NIH, NSF), philanthropic funders (Coefficient/Open Philanthropy), and international bodies (UKRI).
The atlas currently tracks 738 entities and 1,216 connections between them:
- 236 researchers from biosecurity, AI safety, gene editing, and related fields
- 171 universities and research institutions worldwide, with 70 geolocated
- 171 organizations receiving funding for biosecurity and pandemic preparedness work
- 68 research programs funded by DARPA, NIH, NSF, UKRI, and Coefficient
- 57 research papers from four NeurIPS 2025 workshops and key biosecurity reports
- $703M in tracked funding across biosecurity, pandemic preparedness, and AI safety
Nineteen institutions serve as "bridge nodes" — they appear in both the research community and the funding landscape, including Harvard, Stanford, MIT, Oxford, Johns Hopkins, and SecureBio.
Open the live demo and explore:
- Search for any person, institution, or program by name
- Click any node to see its connections, funding totals, and links to source material
- Switch layouts between grouped (by type), force-directed (clustered), and radial views
- Filter by layer to focus on just the workshop community, just the funding landscape, or everything together
- Toggle edge types to highlight specific relationships like funding flows, affiliations, or co-authorship
The visualization is fully self-contained — no server needed, works offline, and runs in any modern browser.
Every connection in the atlas traces back to a verifiable source. The graph was built across 12 sprints, each adding one data source at a time:
- NeurIPS 2025 Workshops: Papers and authors from BioSafe GenAI, MLGenX, GenAI4Health, and AI4Science workshops via the OpenReview API
- Coefficient / Open Philanthropy: 272 bio-relevant grants from their public database, covering $543M in biosecurity, pandemic preparedness, and health R&D
- DARPA & IARPA: Programs including SAFE GENES, PREEMPT, PREPARE, PROPHECY, and Fun GCAT, sourced from official announcements
- NIH: $35M across gene editing, metagenomics, and bio-engineering grants via the NIH RePORTER API
- NSF: 10 matched awards ($9.5M) for key researchers via the NSF Awards API
- UKRI: Pandemic preparedness funding from the UK Gateway to Research API
- Policy & Industry: Biosecurity evaluations from OpenAI, Gryphon Scientific, RAND Corporation, Anthropic, and the NTI AIxBio Forum
- Semantic Scholar: Publication records for key PIs connecting their funded work to the published literature
A 14-point automated quality check runs after every update to catch dangling references, duplicate records, and structural issues.
The full technical roadmap, sprint logs, and schema details are in ROADMAP.md.
To reproduce or extend the atlas:
# Quality check
python scripts/quality_check.py
# Regenerate visualization
python scripts/generate_html_v3.py # → index.html
python scripts/generate_webgl.py # → index_webgl.htmlRequires Python 3.8+, standard library only. All data files are in data/ with raw sources preserved in data/raw/.
Push to GitHub and enable Pages (Settings → Pages → Branch: main, folder: /).
Both index.html and index_webgl.html are self-contained with all data embedded — no CDN dependencies.
Data sourced from OpenReview, Coefficient/Open Philanthropy, NIH RePORTER, NSF Awards, and UKRI Gateway to Research (all public data). Visualization code is MIT licensed.