Respond to scam emails while extracting information about the scammers' tactics, infrastructure, and threat-actor relationships.
- Done with
- ScamBuster
- How
- The AI pretends to be a victim so scammers continue interacting with it, then subtly elicits intelligence that can be shared with security teams, threat-intelligence feeds, and law enforcement.
- Result
- The intended results are intelligence on infrastructure, indicators of compromise, tactics, techniques and procedures, bad-actor profiles, and connections between apparently separate scam campaigns. The transcript does not report results from a live deployment.
- Replaces
- Manual scam-baiting and intelligence collection by individuals, particularly people without the skills to engage scammers safely.
Products from this video
From GPT-Red: Can AI red teams stop prompt injections? by IBM Technology