← Back to timeline
AI THREAT TRACKER / AI-013 · Incident

AI-assisted espionage targets travelers and government personnel

A reported espionage campaign combines stolen data with operational control.

Model labAnthropic
TopicsUnauthorized access
OCCURREDFebruary–August 2026 (exact dates unknown)
PUBLIC ACCOUNT31 Jul 2026
EVIDENCEFirst-party account
About these dates

Microsoft describes activity from February. Anthropic's campaign graphic spans March–August. These bounds do not establish individual intrusion dates or a campaign end. July 31 is the linked earlier public account.

INCIDENT SEVERITY6/10 provisionalSerious impactExternal operations

A central account server was taken over. Control of shared identity infrastructure meets level 6; grave realized harm or prolonged functional loss would be needed for level 7.

Rating criteria →

Sources & attribution

  1. First-party report 10 Sept 2026
    Detecting and countering misuse of AI: September 2026

    Anthropic. The developer is an interested party. Claims and attributions require their stated qualifications.

    Full report (PDF).

  2. First-party report 31 Jul 2026
    CaptiveCrunch: Midnight Blizzard targets travelers worldwide for malware delivery and credential theft

    Microsoft Threat Intelligence. Security-provider observations and attributed analysis. Collaboration credits limit claims of complete independence.