{
 "public_id": "AIR-2025-0013",
 "incident_id": "2025-03-llm-crawlers-foss-infrastructure-dos",
 "date_disclosed": "2025-03-17",
 "date_occurred": "2025-03",
 "title": "LLM training crawlers ignoring robots.txt and spoofing browser identity knocked KDE, SourceHut, GNOME and Fedora infrastructure offline",
 "class": "in_the_wild",
 "attack_surface": "autonomous_ops",
 "initial_vector": "none_autonomous_action",
 "agent_capability_abused": [
  "autonomous_web_crawling",
  "identity_spoofing",
  "proxy_rotation"
 ],
 "target_product": "AI training crawlers operated by Alibaba, Amazon and others",
 "target_vendor": "multiple AI companies",
 "autonomy_level": "fully_autonomous",
 "impact_realized": true,
 "impact_type": [
  "service_disruption",
  "financial_loss"
 ],
 "blast_radius": "KDE GitLab, SourceHut, GNOME and Fedora infrastructure; volunteer-run FOSS hosting generally",
 "cve": [],
 "taxonomy": {
  "owasp_asi": [
   "ASI03"
  ],
  "mitre_atlas": [],
  "labels": [
   "no_adversary",
   "robots_txt_ignored",
   "user_agent_spoofing",
   "residential_proxy_evasion",
   "externalised_cost",
   "autonomous_dos"
  ]
 },
 "defense_present": "robots.txt, user-agent blocking, traffic filtering",
 "defense_bypassed": true,
 "fix_shipped": "none by the crawler operators; projects deployed proof-of-work challenges (Anubis) and country-level blocks",
 "primary_sources": [
  "https://arstechnica.com/ai/2025/03/devs-say-ai-crawlers-dominate-traffic-forcing-blocks-on-entire-countries/",
  "https://incidentdatabase.ai/cite/1001"
 ],
 "verified_quote": "Despite configuring standard defensive measures---adjusting robots.txt, blocking known crawler user-agents, and filtering suspicious traffic---Iaso found that AI crawlers continued evading all attempts to stop them, spoofing user-agents and cycling through residential IP addresses as proxies.",
 "confidence": "high",
 "notes": "An autonomous-agent denial of service with no adversarial intent, which is why it belongs here. The agents actively defeated the consent mechanism — spoofing identity and rotating residential proxies is evasion, whatever the operator intended. Victims are volunteer-run projects absorbing the compute bill. Drew DeVault's post title, 'Please stop externalizing your costs directly into my face', names the pattern. AI Incident Database incident 1001.",
 "risk_category": "reliability",
 "risk_secondary": [
  "security_fraud"
 ],
 "risk_subcategory": "DoS & Sponge / Resource Exhaustion",
 "risk_sub_secondary": [
  "Load & Resilience Failures"
 ],
 "ai_entity": "generative",
 "agency": "agent_acted",
 "agency_debatable": false,
 "agency_reason": "the crawlers autonomously fetched at a rate that took the infrastructure offline",
 "vector_family": "none_autonomous",
 "vector_detail": "none_autonomous_action",
 "guardrail_kind": [
  "unknown"
 ],
 "guardrail_outcome": "silent",
 "xrefs": {
  "aiid": [
   1001
  ],
  "atlas": [],
  "cve": [],
  "ghsa": []
 },
 "primary_asi": "ASI03",
 "agent_framework": "unknown",
 "model_provider": "alibaba",
 "model_name": "",
 "detection_mechanism": "unknown",
 "tool_access": "unknown",
 "financial_loss_band": "unknown",
 "eai_category": "uncontrolled_resource_consumption",
 "eai_domain": "tool_misuse",
 "eai_secondary": [
  "unsafe_actuation"
 ],
 "eai_reason": "subcategory: DoS & Sponge / Resource Exhaustion",
 "asi_curated": [],
 "asi_scope": "in_scope",
 "asi_provenance": "rule",
 "asi_reason": "ASI03 (signal=identity)",
 "submitted_by": "EAI Internal"
}