The issue began with a routine internal query, where an AI-generated response led to unintended data exposure
The incident lasted over two hours and was classified as a high-severity “Sev 1” event by Meta
Despite past AI mishaps, Meta continues aggressive investments to compete with OpenAI, Google, and Anthropic
Meta AI agents reportedly malfunctioned inside the company which led to the exposure of sensitive internal data and user data, as per the report by The Information. It reportedly started with a routine internal query posted by a Meta engineer on the company’s forum seeking help on a technical issue. In response, another employee reportedly deployed an AI agent in order to analyse the problem to which agent generated a reply, which was shared without the original engineer’s approval.
SurveyThe report further mentioned that the AI generated guidance was flawed. Acting on the response, the engineer unintentionally made a large volume of company and user data accessible to other employees who were not allowed to see it. The exposure reportedly lasted for over two hours before being fixed.
Meta reportedly acknowledged the incident and classified it as Sev 1 event, one of the highest levels of severity used by the company to track security risks. This is not the first time Meta has faced issues linked to autonomous AI agents. Previously, Summer Yue shared an experience involving an AI tool from OpenClaw that reportedly ignored instructions and began deleting emails from her Gmail account without confirmation.
Also read: Instagram will not offer end-to-end encryption soon: 5 steps you should take right now to stay safe
Even after such incidents, Meta is doubling down its investments in agentic AI systems. Previously, Meta acquired an AI focused social platform called Moltbook and brought its co-founders into its Meta Superintelligence Labs initiative. The new reports also suggest that Meta is planning to lay off 20 per cent of the workforce globally for AI.
Further, Meta has also made acquisitions such as Manus AI and Limitless, expanding its AI portfolio to rival OpenAI, Google and Anthropic in the race to build next generation AI systems.
Follow Us
Ashish Singh
Ashish Singh is the Chief Copy Editor at Digit. He's been wrangling tech jargon since 2020 (Times Internet, Jagran English '22). When not policing commas, he's likely fueling his gadget habit with coffee, strategising his next virtual race, or plotting a road trip to test the latest in-car tech. He speaks fluent Geek. View Full Profile
