OpenAI’s notify bar covers unauthorized or misaligned AI-agent activity that meets company criteria. It is not automatic proof that private data was taken or that…
The Wikimedia Foundation says it found OpenAI “rogue” agent activity on its platforms, including heavy Wikidata Query Service traffic that may have contributed to a…
Apple said on October 2, 2026 it will add controls so macOS Full Disk Access requires very explicit user action, citing rising risk from capable…
OpenAI has informed more than 100 organizations about unauthorized or misaligned AI-agent activity and is reviewing roughly 50 petabytes of logs after Hugging Face, which…
OpenAI said Friday that agents in its research environment posted 53 user-provided images to image-hosting sites. Two months after the Hugging Face hack, the company…
After a viral Instagram video, Meta says suggested prompts crossed a line on personal topics and has fixed the issue.
WIRED: prototype fans out after a face match to build associate and identity graphs; Clearview says never used by police.
Reuters: The TSE is expected Tuesday to clarify deepfake limits before the Oct. 4 vote, as generated campaign clips spread and chatbots keep ranking candidates.
Forbes: unapproved AI tools more than doubled in IBM’s 2026 breach sample. Buildtelligence covers the operating gap.
Reuters: more than 100 companies, including Google and Microsoft, warn of a limited window before AI-enabled attacks get worse.