{
 "public_id": "AIR-2025-0027",
 "incident_id": "2025-03-anthropic-influence-as-a-service",
 "date_disclosed": "2025-04-23",
 "date_occurred": null,
 "title": "Influence-as-a-service: Claude used as the orchestrator deciding actions for 100+ social media bot personas",
 "class": "in_the_wild",
 "attack_surface": "multi_agent_system",
 "initial_vector": "direct_misuse_as_orchestration_layer",
 "agent_capability_abused": [
  "autonomous_decision_making",
  "persona_management",
  "cross_platform_action_selection"
 ],
 "target_product": "Claude",
 "target_vendor": "Anthropic",
 "autonomy_level": "semi_autonomous",
 "impact_realized": true,
 "impact_type": [
  "influence_operation",
  "platform_manipulation"
 ],
 "blast_radius": "Over 100 bot accounts on X and Facebook engaging tens of thousands of authentic accounts across multiple countries",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI10"
  ],
  "mitre_atlas": [],
  "labels": [
   "orchestration",
   "influence_operation",
   "bot_network"
  ]
 },
 "defense_present": "Anthropic usage policy enforcement",
 "defense_bypassed": true,
 "fix_shipped": "Accounts banned",
 "primary_sources": [
  "https://www.anthropic.com/news/detecting-and-countering-malicious-uses-of-claude-march-2025"
 ],
 "verified_quote": "determine whether social media bot accounts should like, share, comment on, or ignore specific posts",
 "confidence": "high",
 "notes": "Genuinely agentic influence operation: the model was the tactical decision layer, not just a text generator, which is why it is in scope where ordinary IO content generation is not. Anthropic says no content went viral, so realised impact is engagement-level rather than measurable narrative effect. Report page is dated March 2025 in the slug but was published April 2025.",
 "risk_category": "product_misuse",
 "risk_secondary": [],
 "risk_subcategory": "Capability Abuse",
 "risk_sub_secondary": [],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "Claude decided the like/share/comment actions the bot personas then executed",
 "vector_family": "misuse",
 "vector_detail": "direct_misuse_as_orchestration_layer",
 "guardrail_kind": [
  "abuse_detection"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI10",
 "agent_framework": "unknown",
 "model_provider": "anthropic",
 "model_name": "",
 "detection_mechanism": "vendor_disclosure",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "policy_drift",
 "eai_domain": "governance",
 "eai_secondary": [],
 "eai_reason": "subcategory: Capability Abuse",
 "asi_curated": [],
 "asi_scope": "in_scope",
 "asi_provenance": "rule",
 "asi_reason": "ASI10 (adversary-operated agent (vector_family=misuse))",
 "submitted_by": "EAI Internal"
}