Credit: The original article is published here.
Researchers pitted a person against a Claude agent and found that, after a week of texting, the AI chatbot was more effective at creating “exploitable trust” with others.
![[Aggregator] Downloaded image for imported item #160746](https://kcmicro.com/wp-content/uploads/2026/07/ywSwn3oGxXv4PfcRPZmTrc-1280-80-2-800x500.jpg "[Aggregator] Downloaded image for imported item #160746")
![[Aggregator] Downloaded image for imported item #161465](https://kcmicro.com/wp-content/uploads/2026/07/MuH6FdnCJqsnobznT3LSEM-1280-80-1-800x500.jpg "[Aggregator] Downloaded image for imported item #161465")
![[Aggregator] Downloaded image for imported item #161448](https://kcmicro.com/wp-content/uploads/2026/07/Security_OpenAIsHackingDebacleWasaHumanMistake_v1-ezgif.com-video-to-gif-converter-800x500.gif "[Aggregator] Downloaded image for imported item #161448")