Engineered Pathogen Outbreak
A pathogen with clear evidence of deliberate engineering — gain-of-function modifications, synthetic assembly, or otherwise — causes a significant outbreak: at least 10,000 confirmed cases or deaths attributable to the engineered agent.
- Median year
- 2033
- P10 – P90 range
- 2026 – 2050
- Probability ever occurs
- 25%
- Last reviewed
- June 2026
A deliberately engineered pathogen causes a significant outbreak. Whether the origin is state-sponsored bioweapons development, a lab accident involving gain-of-function research, or a non-state actor, the event fundamentally changes the biosecurity and laboratory-safety landscape.
The biosecurity and dual-use research oversight architecture holds — no engineered pathogen crosses the threshold into a significant outbreak. Natural pandemic risk remains; this specific threat is contained.
Where things stand
The risk landscape has shifted structurally. Gene synthesis costs have fallen by orders of magnitude over the past two decades — what required a state-level program in the 1990s can now be attempted by a well-funded non-state actor. AI systems capable of protein function prediction (AlphaFold and successors) lower the barrier further: they enable rational design of novel pathogen variants without extensive wet-lab trial-and-error.
The definitional challenge is significant. The definition requires “clear evidence of deliberate engineering,” which the COVID-19 pandemic illustrated is extremely difficult to establish even with full access to viral genomic sequences. A future outbreak with ambiguous origin may meet the epidemiological threshold (10,000 cases or deaths) without the engineering classification ever being settled. That ambiguity is built into the event definition deliberately — the uncertainty about attribution matters.
The most plausible pathways by source:
- State bioweapons programs: Several states maintain legacy programs and active dual-use R&D. Attribution would be contested and delayed.
- Lab accidents in gain-of-function research: Enhanced pathogens developed for vaccine or therapeutic research could escape containment. The global BSL-4 laboratory network has expanded.
- Non-state actors: Technically harder than state programs, but AI-assisted design and commodity DNA synthesis lower the threshold year by year.
The p_ever of 0.25 reflects that the majority of tail-risk scenarios never materialize. Biosafety and biosecurity infrastructure does constrain risk, even imperfectly. The wide year range (2026–2050) reflects genuine ignorance about timing, not about the direction of technological change.