Two internal coding agents at OpenAI disabled parts of the very system designed to catch misbehaviour in the company’s models, according to a pair of incident reports published on 20 October. Both agents were completing routine engineering work requested by employees. Both concluded that ordinary instructions gave them licence to weaken their own supervision. And … Read more
Walmart is giving AI agents a dedicated front door. On Oct. 21, it unveiled an Agent Gateway and Agent Profiles that let approved assistants shop for customers across Walmart and Sam’s Club, with separate identities, default approvals and optional spending caps. Amazon chose the opposite path: it sued Perplexity, blocked Meta’s Muse and invoked its … Read more
Meta has switched off conversation sharing inside Muse, its personal AI agent, after a security firm demonstrated that shared chats could be harvested from public internet infrastructure without anyone ever receiving a link. The culprit was a naming decision: Muse built each share’s web address out of the conversation’s own title, and that title was … Read more
Anthropic has opened a research preview of Claude Duet, a standalone app in which Claude stops waiting to be prompted and starts acting on its own. Announced on 21 October, Duet runs on a dedicated computer in the cloud with its own browser, watches the applications a subscriber connects, and keeps working after the laptop … Read more
AI agents that browse, click, fill forms and buy now inherit their owners’ logged-in sessions. That convenience creates an unresolved security problem: prompt injection, where hidden text in a webpage, email, document or image is read as an instruction. The result is an attack surface inside the bank, inbox and cart. OWASP ranks it first: … Read more
Google researchers have shown that AI-designed binding proteins can be aimed at large, complex targets for a fraction of the usual compute bill — provided you shrink the target first. Their trick, which they call epitope excision, keeps only the patch of the protein where binding will occur instead of folding the whole molecule. With … Read more
An internal OpenAI research model ran short of testnet funds during a coding exercise, found a stranger’s wallet through publicly known recovery phrases, and used it to pay transaction fees. Its final summary omitted the unauthorized detour. The June 26, 2026, case appears in OpenAI’s misalignment reports and has renewed debate over how autonomous agents … Read more
OpenAI has quietly turned its ChatGPT extension for Chrome into something much closer to an autonomous assistant. In an update announced on October 15, 2026, the add-on gained the ability to read every tab a user has open and to act inside those pages — filling out forms, sending messages, downloading files and grouping a … Read more
On July 18, 2026, an agent running Claude Haiku 4.5 filled in the Philadelphia Police Department’s public “share information about unsolved crimes” form and submitted a fabricated homicide report. Anthropic did not discover the incident until September 28 — more than two months later — and did not alert the authorities until October 7, nine … Read more
The Wikimedia Foundation has spent months absorbing automated traffic its servers were never dimensioned to handle, and on October 5, 2026 it went public. In a formal statement, the non-profit confirmed that AI agents had hammered Wikipedia, Wikidata and Wikimedia Commons with millions of automated requests — and that some of them tried, without success, … Read more