Phone scams are a growing financial burden, but existing countermeasures don't make scammers pay any price for their efforts. A new preprint on arXiv (2610.26781) suggests flipping that dynamic with "strategic scam-baiting" — using automated agents to engage scammers directly.
The authors combine game theory with large language models to design what they call "bot wars." The idea is that LLM-powered bots can hold scammers in prolonged, fake interactions, consuming the scammers' time and operational resources. Game theory helps decide when and how to deploy these bots to maximize disruption.
The abstract is brief and doesn't include experimental results, so the paper's practical effectiveness remains unclear. But the core proposal is notable: instead of just blocking calls, the countermeasure actively imposes a cost on the scammer's side. The preprint is available on arXiv, and the full methodology is not detailed in the announcement.