Meta’s AI smart glasses recordings are reviewed by human contractors training the system, with reports of reviewers encountering deeply personal footage.
Malicious AI browser extensions collected LLM chat histories and browsing data from platforms such as ChatGPT and DeepSeek.