In the wild
Realized harm
North Korean IT workers used Claude to obtain and hold remote jobs at Fortune 500 companies
AIR-2025-0121 AIR record · 2025-08-anthropic-dprk-it-worker-fraud
A real party was affected. This incident caused actual consequences, not a demonstrated capability.
From the supporting source
Operators who cannot otherwise write basic code or communicate professionally in English are now able to pass technical interviews at reputable technology companies.anthropic.com
The record
Disclosed
2025-08-27
Vendor
Anthropic
Product
Claude / Claude Code
Risk domain
Governance › Policy drift
System type
Generative model
Who acted
Human acted on output — Claude gave capability uplift; the workers held the jobs and did the work
Attack surface
Coding agent
Autonomy level
Assistant
Initial vector
Direct misuse for identity and work fraud
Impact type
Fraud, Sanctions evasion, Insider access
Confidence
High
Submitted by
EAI Internal
Crosswalk
- OWASP Agentic Top 10 (2026)
- ASI01 – Agent Goal Hijack
- MITRE ATLAS
- AML.T0010 – AI Supply Chain Compromise
- NIST AI RMF
- GOVERN 1.5
- EU AI Act
- —
- ISO/IEC
- TR 24028; 23894
- AIUC-1
- E. Accountability
- Scope
- agent failure
- Mitigation
- Implement version pinning for models and prompts; Monitor for behavioral drift.
- Taxonomy
- Enkrypt AI agent risk taxonomy · Black-Box Red Teaming of Agentic AI: A Taxonomy-Driven Framework for Automated Risk Discovery
Blast radius
US Fortune 500 technology companies employing fraudulent DPRK remote workers
Capability abused
Defense & remediation
Recorded defense
Claude usage policies and misuse detection
Guardrail kinds
Abuse detection
Guardrail outcome
Silent
Fix shipped
Accounts banned; indicators shared (Aug 2025)
Taxonomy & references
Curator notes
Weakly agentic - the interesting claim is capability uplift removing the training bottleneck for DPRK's IT-worker revenue programme, not tool use. Included because the workers used Claude to actually perform the delivered engineering work, i.e. an agent operating inside victim companies under a false identity. OpenAI reported an equivalent 'Deceptive Employment Scheme' in Feb and June 2025.
Supporting sources
Cite as
Enkrypt AI. "North Korean IT workers used Claude to obtain and hold remote jobs at Fortune 500 companies". AIR, AIR-2025-0121. https://agentincident.vercel.app/i/AIR-2025-0121/ (accessed 2026-10-19).