Elon Musk's weekend boast about his Grok AI agent buying Lego Star Wars kits on X backfired when the screenshot he posted exposed sensitive purchase details. The image included the 10-digit Lego order number, the price paid, delivery and billing addresses, the last four digits of his credit card, and the email address used by the bot. Within hours, an X user named Enderman reportedly used those details with Lego's customer service chatbot to cancel the order, and another user claimed to have changed the delivery address and increased the order to 400 sets using Musk's original payment method. Musk later said his order had not been canceled, and Lego declined to comment on the activity.
The episode illustrates a growing concern for businesses and consumers as AI shopping agents become more common. Musk had suggested on Saturday that Grok could make purchases from a simple prompt, including uploading a picture of a credit card. AI companies such as Meta and OpenAI have also promoted similar agentic shopping features. But the incident shows how a single careless disclosure can turn an AI assistant's convenience into a liability, especially when customer-service bots can be manipulated with the leaked information.
Payment networks are already flagging the risk. Visa warned late last year that bad actors are building legitimate-looking sites designed to fool AI agents into buying fake items, allowing malicious merchants to harvest payment data for unauthorized transactions. Visa said it observed a 25 percent increase in malicious bot-initiated transactions in the second half of 2025, with the figure jumping to 40 percent in the US. The Lego prank was playful, but the underlying exposure points to a serious challenge for the emerging agentic-commerce business model.