A sensitive data breach occurred at Meta, Facebook’s parent company, due to an artificial intelligence (AI) agent. An internal engineer used agentic AI to solve an engineering issue, which inadvertently exposed private data for two hours. Meta maintains that no user data was mishandled, attributing the incident to human error in AI instruction. This breach is part of a troubling trend, with similar occurrences reported at other tech giants like Amazon, highlighting the ongoing risks associated with AI in corporate environments. Security experts note that AI agents lack the comprehensive context human engineers possess, raising concerns about their deployment in sensitive operations. This incident coincides with Meta’s extensive investment in AI despite criticisms surrounding its experimental nature and the environmental impact of data centers. These facilities have faced increased community pushback, with significant capital investments in new data centers being blocked. Meta plans to double its AI investments by 2026 amid growing scrutiny.
Source link
