Bee Cheng Hiang's AI Error Exposes 95,000 Customer Emails

• •Over 95,000 customer email addresses were exposed due to a bad AI prompt.
• •The incident was classified as Singapore's first AI-related data breach.
• •Bee Cheng Hiang has implemented new policies to prevent future occurrences.

In Singapore's first AI-related data breach, over 95,000 email addresses of Bee Cheng Hiang customers were exposed due to a bad prompt used by an employee in an AI tool. The incident occurred in April when the employee generated a Python script for bulk marketing emails, causing all recipients' addresses to be visible. The Personal Data Protection Commission (PDPC) confirmed that the breach was due to human error, not a malfunction of the AI tool. There is no evidence of further misuse of the exposed email addresses. Bee Cheng Hiang has since implemented stronger data protection measures, including double-verification for bulk emails. The PDPC noted that the company lacked a governance framework for AI tool usage, which contributed to the incident. The breach was reported to the PDPC on April 27, shortly after the emails were sent out on April 25.