WorkOS says an AI-powered hacking agent that breached its internal sales application and accessed customer information was operated by a security researcher who claims the tool escaped its testing environment and unintentionally attacked the company. [...]
BleepingComputer
The UK is betting on domestic AI and cyber investment to reduce reliance on US tech — but security leaders question whether the funding is too fragmented and too late.
Dark Reading
Security records are increasingly being used by vendors to train their models, creating new risks for enterprises.
Dark Reading
Deep-Live-Cam est un outil open source capable de remplacer un visage en direct lors d'une visio (Teams, Zoom, etc.) à partir d'une seule photo. Le post Deep-Live-Cam : cet outil open source de deepfake change votre visage à partir d’une photo a été publié sur IT-Connect.
IT-Connect
OpenAI on Thursday unveiled GPT‑6 Cyber, which it described as a major step forward in security research that brings together stronger cybersecurity capabilities with a step-change improvement in model alignment. The new model is "more capable at finding and fixing vulnerabilities, and far more reliable staying within the boundaries researchers set," OpenAI said. It has classified the model as
The Hacker News
The U.S. is working on extraditing three suspects from Europe in a case involving more than $100 million in fraud from CEO impersonations.
The Record
Rated Critical for cybersecurity, OpenAI’s new model beats GPT-5.6 Cyber at exploitation while staying in scope more reliably. The post OpenAI Says New GPT-6 Cyber Model Hunts Flaws Without Crossing the Line appeared first on SecurityWeek.
SecurityWeek
Grâce à son mécanisme de sécurité de la mémoire et ses performances, Rust gagne en popularité pour la modernisation de back-end. Sa courbe d’apprentissage abrupte en a d’abord fait le langage des fournisseurs. L’IA agentique chamboule ce schéma. Mais attention : s’il est bien plus facile de réécrire du code en Rust, une expertise approfondie en matière d’architecture demeure essentielle.
LeMagIT
ARTEX, a Chinese open-source agentic penetration-testing platform, to run autonomous attacks. The operators set a goal and then mostly check progress. Researchers identified 2,557 systems that host or hosted ARTEX panels, and 1,608 were live and public. Ten misconfigured instances exposed their full databases. Those databases show the models, prompts and actions behind the campaigns. […] The post
GBHackers
After its sandbox proxy failed, an internal OpenAI model crafted raw packets and probed AWS services trying to reach the internet. The post OpenAI Model Wrote Its Own Network Code to Get Around Sandbox Restrictions appeared first on SecurityWeek.
SecurityWeek
A design flaw in Anthropic’s Claude Dispatch feature let a malicious browser extension run commands on the operating system, even on a different computer. Researchers at Forever Security named the attack ThisBash. They found that Claude Dispatch accepted instructions from an ordinary logged-in claude.ai session. An extension that could control that site could therefore send […] The post Claude AI
GBHackers
Britain's new Labour-led government is making a concerted push to get ahead of the AI risks to its national networks and the industrial complex.
Dark Reading
Les acteurs malveillants ciblent et exploitent depuis longtemps les vulnérabilités des navigateurs. La généralisation de l’IA accroît ce risque et oblige les RSSI à réexaminer leurs options en matière de sécurité du navigateur.
LeMagIT
Mistral AI poursuit ses emplettes. Cette fois-ci c’est SkipLabs, une startup spécialisée dans la programmation réactive qui donne naissance à une nouvelle équipe. Son objectif : créer des outils pour valider et tester le code généré, le fiabiliser, avant de le soumettre aux développeurs.
LeMagIT
Attackers using the ARTEX agentic AI framework to discover exposed GitHub credentials and rapidly compromise cloud environments. In one intrusion, an AI agent progressed from searching GitHub for application secrets to exfiltrating Azure Key Vault data in just five minutes. The activity compresses a familiar attack chain: finding leaked credentials, validating access, mapping resources, and […] Th
GBHackers
OpenAI lance GPT-6 Cyber, réservé à Daybreak Red : 948 exploits Linux, 81 failles graves dans Firefox et un modèle entraîné à rester dans son périmètre. Le post OpenAI lance GPT-6 Cyber : son modèle IA taillé pour trouver et exploiter des failles a été publié sur IT-Connect.
IT-Connect
The trio, who have nicknames that include “Rocky Balboa,” “Superman Cash” and “Fat Boy,” often targeted the companies’ accounting departments. The post 3 Charged in Scheme to Use AI to Impersonate CEOs and Attorneys to Steal $129 Million appeared first on SecurityWeek.
SecurityWeek
Les agents d’IA continuent pour l’essentiel de jouer un rôle consultatif auprès des équipes NetOps, qui adoptent une approche de type « faire confiance, mais vérifier ». Cependant, les systèmes autonomes gagnent du terrain à mesure que la complexité des réseaux augmente.
LeMagIT
Hackers are using fake OpenAI, SpaceX, and Anthropic event invitations to lure employees into phishing attacks, using two different phishing-as-a-service platforms to steal Microsoft and Google account credentials and authentication sessions. [...]
BleepingComputer
OpenAI has disclosed another AI alignment incident where GPT-6 Astra was found to sometimes quietly withhold help on safety research. [...]
BleepingComputer
An unreleased OpenAI cyber model found and exploited 948 distinct Linux kernel privilege escalation bugs for about $1.2 million. The post AI Agents Got Root on Linux via 948 Kernel Flaws in OpenAI Research appeared first on SecurityWeek.
SecurityWeek
Supprimez les pubs de l'accueil de votre box Android TV avec CleanYourTV et Claude Code, ou une app gratuite : ADB, prérequis, thèmes et précautions. Le post CleanYourTV : comment nettoyer une box Android TV avec l’IA (sans root) a été publié sur IT-Connect.
IT-Connect
The ungated resource explains how security data pipelines reduce SIEM ingestion costs, preserve detection coverage, and prepare the data layer for AI-driven security operations.
Dark Reading
Cybersecurity researchers have disclosed details of a now-patched vulnerability that could be exploited to bypass Lockdown Mode protections in OpenAI ChatGPT. Lockdown Mode was introduced by the artificial intelligence (AI) company in June 2026 as an optional setting to minimize data exfiltration risks arising from prompt injections. "The flaw could enable silent chat data theft, phishing,
The Hacker News
Lors de la GTC à Berlin, SUSE a été mis en avant lors du keynote d’ouverture de Nvidia. Le spécialiste des distributions Linux et de Kubernetes devient un des relais des architectures de référence du géant fabless en Europe, qui entend s’en servir pour accélérer la mise en production des projets d’IA. Et un prétexte pour évincer progressivement vSphere.
LeMagIT
