AI models will be capable enough for terrorist groups to create novel pathogens (near future).

AI models will be capable enough for terrorist groups to create novel pathogens (near future). | we're not that far away from a world where there are incredibly capable open source models that are very good at biology. And the need for society to be resilient to terrorist groups using these models to try to create novel pathogens is like that's no longer a theoretical thing or it's not going to be for much longer.

we're not that far away from a world where there are incredibly capable open source models that are very good at biology. And the need for society to be resilient to terrorist groups using these models to try to create novel pathogens is like that's no longer a theoretical thing or it's not going to be for much longer.

Beneficiaries (23)

Adverse (6)

{
  "nia": false,
  "url": "https://www.youtube.com/watch?v=cFI-SqnvQK8",
  "mode": "CITED_PREDICTION",
  "role": "Cited-Executive",
  "context": "we're not that far away from a world where there are incredibly capable open source models that are very good at biology... terrorist groups using these models to try to create novel pathogens",
  "to_year": 2027,
  "cited_by": "Peter Diamandis",
  "verbatim": "we're not that far away from a world where there are incredibly capable open source models that are very good at biology. And the need for society to be resilient to terrorist groups using these models to try to create novel pathogens is like that's no longer a theoretical thing or it's not going to be for much longer.",
  "conv_cues": "not that far away; no longer theoretical",
  "direction": "HAPPEN",
  "from_year": 2026,
  "timeframe": "Near-term",
  "conv_level": "HIGH",
  "milestones": [
    {
      "kind": "llm_pre_event",
      "label": "OpenAI April 2025 risk assessment flags 'high' biorisk capability",
      "notes": "HIT — OpenAI formally flagged 'high risks' substantially increasing likelihood of bioterrorist attacks. Validates Altman's 'no longer theoretical' framing.",
      "source": "ComputerUser — OpenAI says its next AI might be powerful enough to aid bioweapons",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -10,
      "source_id": null,
      "confidence": 0.95,
      "source_url": "https://computeruser.com/openai-says-its-next-ai-might-be-powerful-enough-to-aid-bioweapons",
      "expected_date": "2025-04-30",
      "observed_date": "2025-04-30",
      "research_origin": "deep_research",
      "measurement_criterion": "OpenAI publishes formal risk assessment concluding next models meet 'high' bio-risk threshold under their Preparedness Framework"
    },
    {
      "kind": "llm_pre_event",
      "label": "Anthropic Claude 3.7 Sonnet shows concerning bioweapon assistance",
      "notes": "HIT — Anthropic's own published assessment confirms Altman's claim direction.",
      "source": "Anthropic — Biorisk red team report",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -9,
      "source_id": null,
      "confidence": 0.95,
      "source_url": "https://red.anthropic.com/2025/biorisk/",
      "expected_date": "2025-07-17",
      "observed_date": "2025-02-15",
      "research_origin": "deep_research",
      "expected_date_range": {
        "to": "2025-12-31",
        "from": "2025-02-01"
      },
      "measurement_criterion": "Anthropic publishes evaluation showing Claude 3.7 Sonnet 'demonstrates concerning improvements' in supporting biological weapons development"
    },
    {
      "kind": "llm_pre_event",
      "label": "AI labs deploy CBRN-content training data filtration",
      "notes": "HIT — Industry response to validated risk; ~5% inference compute spent on bioweapons classifiers per Amodei.",
      "source": "Anthropic alignment team blog (Aug 2025)",
      "status": "hit",
      "weight": 0.4,
      "ordinal": -8,
      "source_id": null,
      "confidence": 0.95,
      "source_url": "https://red.anthropic.com/2025/biorisk/",
      "expected_date": "2025-12-15",
      "observed_date": "2025-08-15",
      "research_origin": "deep_research",
      "expected_date_range": {
        "to": "2026-06-30",
        "from": "2025-06-01"
      },
      "measurement_criterion": "Anthropic / OpenAI publicly disclose deploying classifier-based dataset filtration to scrub CBRN content from training"
    },
    {
      "kind": "llm_post_event",
      "label": "Open-source model with strong biology capability released",
      "notes": "Cascade — Open-source models present unique unfilterable risk per CSIS analysis.",
      "source": "Hugging Face leaderboards; major model release announcements",
      "status": "pending",
      "weight": 0.4,
      "ordinal": -7,
      "source_id": null,
      "confidence": 0.85,
      "expected_date": "2026-03-17",
      "research_origin": "training",
  
... (truncated)