What happened

Developers using Cursor experienced unexpected logouts when switching between devices - a disruption for users who rely on multi-device workflows. 

Seeking answers, a user contacted Cursor’s support and received an email from “Sam,” an AI-powered support agent, stating that the logouts were “expected behavior” under a new policy limiting each subscription to a single device. 

In reality, no such policy existed; the AI had hallucinated the explanation, presenting it as an official company rule.

The fabricated policy quickly spread on forums like Reddit and Hacker News, with users expressing frustration and some publicly canceling their subscriptions. 

The incident escalated as more users believed the AI’s response was a legitimate policy change, threatening Cursor’s reputation and customer base.

Why it happened

The incident appears to have been caused by an AI “hallucination.” Instead of admitting uncertainty, the AI support agent invented a policy to explain the unexpected logouts, prioritising a confident and coherent response over accuracy. 

The underlying technical issue was a backend change intended to improve session security, which inadvertently caused session invalidation across devices. 

However, the AI’s fabricated explanation misled users and amplified the problem.