Cursor, an AI-powered code editor, faced backlash this week after its support chatbot provided inaccurate information. The issue began when a developer was logged out while switching devices.

The AI support agent, “Sam,” incorrectly claimed it was due to a new policy limiting subscriptions to one device.

However, the company later clarified that no such policy existed and blamed the error on a backend update that invalidated some user sessions.

The false policy announcement frustrated users, leading to complaints on Reddit and subscription cancellations.

In response, Cursor apologized, confirmed that users can access the platform on multiple devices, and refunded the affected user. They also revised their support practices by labeling AI-generated responses.

🔗 Source: Wired

🧠 Food for thought

1️⃣ AI hallucinations in customer service follow a costly pattern

Cursor’s experience reflects a mounting trend of corporate AI systems inventing policies and causing business damage when deployed without sufficient safeguards.

The incident joins a growing list of similar high-profile AI failures, including Air Canada being legally forced to honor a refund policy fabricated by its chatbot. These cases highlight a systemic problem: when AI systems invent information rather than acknowledging uncertainty, the consequences extend beyond frustrated customers to include legal liability, financial losses, and damaged brand reputation.

A particularly troubling aspect is that many businesses don’t recognize this pattern until they experience it themselves, despite 92% of companies now using AI for customer interactions 1.

The business impact can be immediate and measurable. Cursor experienced subscription cancellations and public backlash within hours of the fabricated policy being shared.

2️⃣ The transparency gap undermines customer trust in AI

Cursor’s failure to disclose that “Sam” was an AI agent represents a broader industry problem with transparency that directly impacts customer trust.

Recent surveys show 55% of customers already feel frustrated with AI chatbots asking excessive questions and 47% struggle to get accurate information. These factors increase the backlash when AI systems are discovered to be making up information 2.

The confusion created when AI systems present as humans (with names like “Sam”) creates a fundamental deception that amplifies negative reactions when hallucinations occur. Many Cursor users specifically cited this misrepresentation in their complaints.

Recent Cursor developments