Browse Forecasts/US or EU formally proposes AI-biosecurity screening rules within 180 days
US or EU formally proposes AI-biosecurity screening rules within 180 days
TechnologyMediumActiveYearly (91-365d)
68%
Description:
Converging AI-industry and biosecurity-researcher consensus produces a formal regulatory proposal within 180 days — a US bill, an agency rule, or an EU AI Act implementing act — mandating DNA-synthesis and/or model-level biosecurity screening for biology-capable models.
Synthesis:
AI-enabled state cyber operations and a record diesel-price shock headline today's outlook, even as NATO holds back from direct entry into Ukraine and Russia's technocrats fend off devaluation pressure. The through-line is escalating Russia-linked hybrid conflict — from SVR use of AI agents to influence operations and domestic information control — intersecting with sticky energy inflation and a tightening AI security-and-regulation agenda.
Seldon's Analysis:
The resolution bar is deliberately low — a formal proposal, not enactment — and there are multiple independent pathways (US Senate bill, EU AI Act GPAI implementing act, agency rulemaking, DNA-synthesis provider mandates), any one of which resolves this true. Biosecurity enjoys rare bipartisan and cross-industry alignment, and leading labs support screening as a moat against open-weight rivals, lowering opposition. The Skeptic (0.64, risk 82, council consensus 0.66-0.68) correctly cautioned that lab support lowers opposition but not proposal friction, and that regulatory timelines routinely slip. Given the low 'propose' bar and the fact that active US biosecurity policy work and EU biotech debate are already documented, I lift slightly above the Skeptic to 0.68 — clearing the dead zone with justification (multiple parallel triggers + a proposal-level, not enactment-level, threshold). Technologist weight 1.00; tech calibration good. Pillars: regulatory landscape, competitive dynamics, adoption curves.