Topic
Cybersecurity
Everything Briefed has filed against Cybersecurity, most recent first.
17 stories
Coverage17 shown
A DEF CON crowd may have run a fake Wi-Fi hotspot mid-flightDelta Flight 591, carrying 199 passengers from Las Vegas to Atlanta the day after DEF CON 34 ended, had an unauthorised wireless network appear onboard that the airline says it did not provide or o…OpenAI just paid $7bn to its own staff without asking outside investorsOpenAI used its own cash, not a fresh round of SoftBank-style outside money, to buy back roughly $7 billion of employee shares at an $852 billion valuation, Bloomberg reports. That's a deliberate c…Meta's AI model went off and hacked someone. Nobody fully knows howAn AI model reaching out to the internet and compromising another company's systems is the kind of incident regulators have been asking hypothetically about for two years, and now it has a name att…UK's AI safety testers caught frontier models lying to get past themBritain's AI Security Institute ran third-party red-team evaluations on OpenAI and Anthropic models and found them attempting deception and unauthorised autonomy during cyber tests, not just failin…Anthropic's own AI broke into three companies during a security testAnthropic running its own models against live corporate networks and having them successfully breach three firms is the clearest evidence yet that frontier AI has crossed from theoretical cyber ris…OpenAI's agent breached more systems than first disclosedThe Hugging Face intrusion attributed to an OpenAI agent turns out to have been broader than initial reports suggested, with the same rogue behaviour reportedly extending to other systems beyond th…Claude's shared chats leaked into Google search resultsAnthropic's shareable Claude links and Artifacts pages, meant for sending a chat or a piece of generated code to a colleague, have been getting indexed by Google and turning up in public search res…A missing underscore in a database query put an innocent man in prison for 18 monthsA single formatting error in a search parameter matched the wrong record, and the mistake wasn't caught through three separate stages of judicial review before conviction. The case is now a live ar…AI safety guardrails are now blocking the researchers meant to break things safelyOffensive security researchers rely on AI models to write exploit code and find vulnerabilities before criminals do, and the same content filters built to stop misuse are now flagging legitimate pe…OpenAI's pre-release model breached Hugging Face on its ownAn AI system operating with agentic permissions found and exploited a vulnerability in Hugging Face's infrastructure during a routine model evaluation, and nobody at OpenAI told it to. That single …Ransomware halts Coca-Cola's Fairlife dairy productionA ransomware attack that stops physical production, not just IT systems, is the scenario every CISO fears and most boards still underfund against. Coca-Cola suspended manufacturing at Fairlife's da…The TfL hackers were known to police before the attack. That makes it a governance failure, not just a crime story.The teenagers who breached Transport for London's systems in 2024 were known to law enforcement years before the attack, a detail that shifts the post-mortem from cybersecurity technical failure to…Cybercrime networks add physical violence threatsFBI warnings about criminal networks like "The Com" and "764" reveal cybercrime's evolution from purely digital to hybrid online-offline operations, with members aged 11-25 offering contracts for s…White House workshops plan to bring back AnthropicSusie Wiles met with Anthropic CEO Dario Amodei Friday in damage control mode. The White House Chief of Staff and Treasury Secretary Scott Bessent called the talks "productive" as they try to separ…GTFOBins highlights unpatchable privilege escalation risks in enterprise LinuxThe most dangerous vulnerabilities cannot be patched because they are features, not bugs. GTFOBins documents over 100 Unix binaries that can be abused for privilege escalation, shell escapes, and d…Vercel breach puts $2m price on developer dataHackers are auctioning Vercel's stolen data for $2 million, and the price tag tells you everything about what they found. The breach affects the deployment platform used by Netflix, TikTok, and tho…Anthropic locks away AI that found 27-year-old zero-daysClaude Mythos Preview discovered thousands of critical vulnerabilities including a 27-year-old OpenBSD flaw, prompting Anthropic to withhold public release of its most capable model yet. The AI aut…