← Trends

Autonomous Agents Move Into Cyber Offense

AI agents increasingly run end-to-end intrusions, chaining supply-chain footholds into privilege escalation and credential theft at machine speed, outpacing human and current automated defenses.

strengthening · confidence 100 · +2 7d · 0 30d · Medium term (3-9 months) · tracking since July 21, 2026 · updated September 14, 2026

Sign in to get threshold and movement alerts for this trend.

Score history

Daily conviction score, 0 to 100. Higher means the thesis is more strongly corroborated.

Sep 13 · 97Sep 14 · 100

Now 100 · +3 since Sep 13 · ranged 97 to 100

Showing the last few days. Unlock full score history.

Why the conviction moved

  • Sep 14
    Strengthened +3

    Anthropic disclosed disrupting Claude misuse across seven harm categories over nine months, spanning influence, fraud and intrusion-adjacent operations. Lab-side enforcement telemetry, rather than victim reporting, is becoming the primary public record of agentic offensive activity — which means the observed volume tracks detection capacity, not the underlying rate.

  • Sep 13
    Strengthened +5

    Anthropic's threat report documents a Russia-linked group using Claude Code to build autonomous drone targeting software, alongside industrial-scale automated persona operations. This extends agentic misuse past intrusion tradecraft into autonomous kinetic targeting built by a coding agent, widening the harm surface the thesis tracks.

Showing the last 2 days. Unlock the full record.

Source trail

  • Supporting · September 14, 2026

    Anthropic Says It Disrupted Claude Misuse Across Seven Harm Categories in Nine Months

    Anthropic disclosed disrupting Claude misuse across seven harm categories over nine months, spanning influence, fraud and intrusion-adjacent operations. Lab-side enforcement telemetry, rather than victim reporting, is becoming the primary public record of agentic offensive activity — which means the observed volume tracks detection capacity, not the underlying rate.

    Anthropic
  • Supporting · September 13, 2026

    Anthropic Says Claude Models Can No Longer Be Assumed Below the Bioweapons Assistance Threshold

    Anthropic's threat report documents a Russia-linked group using Claude Code to build autonomous drone targeting software, alongside industrial-scale automated persona operations. This extends agentic misuse past intrusion tradecraft into autonomous kinetic targeting built by a coding agent, widening the harm surface the thesis tracks.

    Anthropic News

Unlock full source trail, score history, and daily updates.

42 more sources in the full trail.

Unlock Trends

Affected regions & assets

RegionsGlobal

Townsquare

Argue the thesis in Townsquare.