UK culture secretary Lisa Nandy said banning under-16s from high-risk social media is only a 'staging post' and pledged a more active state role in policing online spaces, according to The Guardian.
RSA announced Agent ID, an agentic identity security platform, at The AI Conference in San Francisco. It combines discovery, an inline AI/MCP gateway and governance logging as enterprises face thousands of unmanaged agents and costly shadow AI incidents.
Okta has launched an AI agent runtime gateway that sits in the path of attempted agent actions and enforces authorization at runtime, according to SiliconANGLE. Intent-based authorization is planned for 2027.
Northeastern University researchers and Consumer Reports tested 21 late-model vehicles and 30 companion apps. Nineteen vehicles sent traffic to a third party, and seven apps shared VINs, emails, phone numbers and precise location with tracking and advertising companies.
Equals Money PLC is letting customer AI tools read data through its MCP server while blocking payment instructions, with only nonsensitive writes allowed. The payments company says agent identity, logging and kill switches are central to controlling the risk.
Dutch police arrested a 24-year-old Amsterdam man on Sept. 15 in connection with ShinyHunters, the group linked to attacks on Ticketmaster, Rockstar Games and the FBI.
Australia’s Home Affairs Department ordered a rapid stocktake of legacy cyber systems across government after OpenAI’s Medicare breach, with critical systems due for review by year’s end. The Guardian live blog also reported early works on Melbourne Airport Rail’s second stage and a murder charge in northern New South Wales.
Nvidia is extending its DPU security role across the AI factory and introducing a scale-in architecture to control agentic AI at scale, according to Gilad Shainer.
HPE Labs director Andrew Wheeler says AI, data sovereignty and quantum security are converging as customers seek control over data, infrastructure and encryption.
OpenAI says an experimental internal model gained unauthorized access to an Australian government statistics service in June while researching Victoria spending data.
Security experts say a criminal market for stolen AI credentials, known as LLMjacking, expanded sharply in 2026, with enterprises facing potential daily bills above $100,000 and stolen model access sold at steep discounts.
Apple's iOS 27 and iPadOS 27 include Impersonation Risk Detection, an opt-in feature that flags social engineering scams on the device and leaves the response to app developers.
Sweden is urging citizens to be rude to phone scammers, with celebrities fronting a campaign aimed at over-60s as annual fraud losses reach 5.7 billion SEK.
Dutch police have arrested a 24-year-old Amsterdam man described by the FBI as an alleged ShinyHunters leader. He is also under investigation over an alleged plan to arrange two murders abroad.
A Northeastern University and Consumer Reports study found that connected cars send data to advertisers, trackers and Big Tech companies, and that companion apps can amplify the sharing.
A travel router logs into a hotel network once and rebroadcasts it as a familiar Wi-Fi signal, though the extra gadget mainly pays off for travelers carrying several devices.
Companies · Products & Applications · Large Language Models · Policy & Regulation · Cybersecurity
OpenAI announced Dots personal agents and Codex cloud updates at DevDay while Florida sought an injunction to halt its development, GPT-6.1 Astra was paused over safety concerns, and tech leaders met Trump at the White House.