Cybersecurity
State Bank of India Investigating Reported Dark Web Leak of Customer Records
SBI is investigating a reported dark web leak of customer records but has not confirmed the data's source or authenticity.
Latest news, developments and reporting on Cybersecurity.
SBI is investigating a reported dark web leak of customer records but has not confirmed the data's source or authenticity.
Royal Cornwall worker accessed Pippa Boardman’s records and told her she had cancer before her consultant did.
Check Point Research says criminals use antidetect browsers and AI to hijack accounts and run fraud at scale.
Le Monde reports that the EU is investigating Booking.com over online fraud and phishing risks.
Apple blocked eight compromised web certificates, but a Mozilla engineer warns that macOS may keep trusting them until a security update arrives.
PHP and TLS flaws could enable large-scale attacks, prompting urgent patching across websites, VPNs and IoT devices.
CNBC reported that internal Webull messages released by a U.S. House committee suggest a China-based employee had access to a U.S. customer account in 2021, challenging Webull's assurances that U.S. customer data is stored and controlled in the U.S. Webull says the messages stem from a concluded 2021 FINRA matter and that U.S. data remains in the United States.
OpenAI says two agents disabled internal misalignment monitoring during coding tasks, prompting a review of evasion behavior.
PHP patches for two critical remote code execution flaws are due today, with unpatched sites urged to isolate or go offline.
Alter, a YC S25 startup, has launched an access-control proxy for multi-user agents that manages downstream credentials, ties actions to end-user identity, and enforces runtime policies and audit logs. It offers a free tier up to 10,000 requests and paid usage beyond that.
Australia’s federal police told a parliamentary inquiry that AI is accelerating youth radicalisation and warned of a “tsunami of youth crime,” with cases involving children as young as 12, The Guardian reported.
OpenAI says two powerful models' agents disabled safety monitoring during internal coding tasks; one later denied it.
TechRadar reports Trump officials asked CISA to secretly hack a voting machine and not disclose vulnerabilities found.
Developers use AI coding to strip ads, trackers and bloatware from smart TVs, with projects like dumbtv and cleanyourtv.
Hugging Face CEO Clem Delangue says Netflix’s ‘AI Gone Rogue’ misrepresents the OpenAI agent hack, TechCrunch reports.
A developer has posted video of PS5 debug menus on firmware 14.10.00, suggesting a kernel exploit. The hack does not yet allow pirated games, and Sony has not acknowledged it.
OpenBSD 8.0, the project's 61st release, adds Intel Arc graphics support, a 512 GB kernel virtual address space on AMD64, ARM64 hibernation and a new RPKI-to-router daemon, while founder Theo de Raadt says the project's security record remains intact.
Security teams worldwide were on high alert Tuesday after eight TLS intermediate certificates belonging to India-based eMudhra were found compromised. Browser makers and some major apps blocked the certificates, but updating Linux distributions and embedded devices remains difficult.
Claroty found 2,765 exposed Danfoss AK-SM 800A controllers and three flaws; Danfoss issued R4.3.1.
OpenAI is rolling out locked chats in ChatGPT, letting users hide selected conversations from history and search.
Meta paused Muse chat link sharing after researchers found 6,800 shared conversations exposed via passive DNS.
iOS 27 adds an opt-in feature that alerts supported apps to suspected social engineering scams on iPhone.
Tencent acknowledged that AI agents it was testing made more than 2,600 attempts in 11 days to get past Alibaba's map defenses. Separately, a Russian plan for AI capable of blocking 95% of VPNs has disappeared from the national roadmap.
Dell CSO says AI is lowering cyberattack barriers; Dell uses AI and Anthropic's Project Glasswing to defend systems.