{
  "id": "wormgpt-underground-2023",
  "name": "WormGPT underground malicious LLM",
  "date_reported": "2023-07-13",
  "researcher": "Check Point / Wired / industry",
  "org": "Check Point Research / underground market reporting",
  "sources": [
    {
      "url": "https://research.checkpoint.com/2023/opwnai-cybercriminals-starting-to-use-chatgpt/"
    },
    {
      "url": "https://www.wired.com/story/chatgpt-scams-fraudgpt-wormgpt-crime/"
    }
  ],
  "magicsword_coverage": [],
  "borderline": true,
  "borderline_note": "Underground tool offering; Check Point assessed WormGPT likely real but not clearly superior to jailbroken mainstream models.",
  "ai_role": "attacker_uses_ai_as_operator",
  "ai_role_secondary": null,
  "ai_role_detail": null,
  "technique_summary": "GPT-J-based unrestricted LLM marketed on dark-web forums for phishing and malware assistance without ethical filters.",
  "initial_access": "unknown",
  "ai_tooling_and_models": [
    "WormGPT (GPT-J based)"
  ],
  "other_tooling": [],
  "victims_and_sectors": "unknown",
  "regions": [
    "unknown"
  ],
  "attribution_or_cluster": "underground cybercrime marketplace",
  "scale_and_metrics": "unknown",
  "prevention_angle": "Classic phishing/malware controls; monitor for AI-polished lure language",
  "related_magicsword_projects": [],
  "source_notes": null,
  "last_reviewed": "2026-10-20",
  "attack_flow": {
    "nodes": [
      {
        "id": "step-1",
        "title": "Underground seller markets WormGPT",
        "description": "A GPT-J-based tool is offered through cybercrime forums.",
        "kind": "access",
        "source_urls": [
          "https://research.checkpoint.com/2023/opwnai-cybercriminals-starting-to-use-chatgpt/",
          "https://www.wired.com/story/chatgpt-scams-fraudgpt-wormgpt-crime/"
        ]
      },
      {
        "id": "step-2",
        "title": "Unrestricted assistance is advertised",
        "description": "The offering claims support for phishing and malware without safety filters.",
        "kind": "ai",
        "source_urls": [
          "https://research.checkpoint.com/2023/opwnai-cybercriminals-starting-to-use-chatgpt/",
          "https://www.wired.com/story/chatgpt-scams-fraudgpt-wormgpt-crime/"
        ]
      },
      {
        "id": "step-3",
        "title": "Individual victim use is not established",
        "description": "The seed documents an offering and researcher assessment, not a confirmed exploit chain.",
        "kind": "unknown",
        "source_urls": [
          "https://research.checkpoint.com/2023/opwnai-cybercriminals-starting-to-use-chatgpt/",
          "https://www.wired.com/story/chatgpt-scams-fraudgpt-wormgpt-crime/"
        ]
      }
    ],
    "edges": [
      {
        "source": "step-1",
        "target": "step-2"
      },
      {
        "source": "step-2",
        "target": "step-3"
      }
    ]
  },
  "category": "adversary-use-of-ai",
  "attribution": {
    "actor_name": "underground cybercrime marketplace",
    "motivation": "financially-motivated"
  },
  "ai_involvement_confidence": "unverified",
  "detections": [
    {
      "name": "Review attack-specific indicators",
      "format": "guidance",
      "notes": "Monitor for the techniques described in WormGPT underground malicious LLM. Refer to sources for specific IOCs and TTPs."
    }
  ]
}