evaluation-task · Documentary review

Make Me Pay

Evaluates an AI models' susceptibility to social engineering attacks by testing whether a "con-artist" model can persuade a "mark" model to donate money through manipulation and persuasion tactics.

Open interactive record →
OriginOpenAI EvalsTopicsHuman influence & agencyStatusReviewed

Can support

Model-to-model success at obtaining the benchmark's simulated financial outcome from the specified model counterpart under controlled dialogue conditions.

Cannot support by itself

Human susceptibility to fraud, actual financial loss, prevalence of scams, lawful or covert deployment, or real-world victimization.

Decision use

Best used for

Studying goal-directed social-influence strategies and model-to-model interaction under a fixed simulated task.

Not enough for

Claims that a model can defraud people, cause financial harm, or automate effective scams in deployment.

Original sources