OpenAI engineers are building a custom Windows sandbox for their Codex coding agent after discovering that existing Windows security tools like AppContainer and Windows Sandbox fail to meet the strict isolation and flexibility demands of open-ended developer workflows, leading to a new elevated solution using dedicated local users, Windows Firewall …
Mistral AI is in talks with European banks to deploy a cybersecurity-focused AI model as an alternative to Anthropic's Mythos, amid growing concerns that AI-powered tools are exposing financial institutions to unprecedented security vulnerabilities.
AI-powered coding is fueling a security crisis as Experian data reveals 40% of last year's 5,000 data breaches were AI-driven, prompting Synthesia to release a six-step automated code security review tool that scans for vulnerabilities at under $4 per review, while experts urge companies to act now as powerful models …
AI-powered fraud now accounts for 40% of data breaches, driving global losses past $534 billion annually, as tools like FraudGPT fuel synthetic identities and deepfakes — forcing companies like OpenAI and TransUnion into an urgent AI-versus-AI arms race to fight back.
Microsoft's AI-driven vulnerability system MDASH has discovered 16 previously unknown Windows flaws, including four critical remote code execution vulnerabilities now patched, using over 100 specialized AI agents to scan source code and top a major cybersecurity benchmark.
Microsoft's new AI security system MDASH, orchestrating over 100 specialized agents, achieves an industry-leading 88.45% on a real-world vulnerability benchmark and directly uncovers 16 new Windows CVEs — including four Critical remote code execution flaws — patched in today's Patch Tuesday release.
AI cyber capabilities are doubling every 4.7 months, with two new models — Claude Mythos Preview and GPT-5.5 — shattering existing benchmarks and triggering urgent security alarms, as experts warn both attackers and defenders face a rapidly shifting threat landscape demanding immediate action.
ServiceNow expands its AI control tower with behavior monitoring and ROI tracking as managers face surging workloads from overseeing autonomous AI agents, while experts warn that unmonitored token usage and data leakage risks are creating serious financial and security threats for unprepared organizations.
Security researchers expose over 5,000 unsecured AI-built apps leaking sensitive real-world data — including hospital records and financial documents — as AI coding platforms deflect responsibility, leaving non-technical users unknowingly exposing critical information to the public.
Mozilla leverages AI models, including Claude Mythos Preview, to uncover 271 hidden Firefox security bugs—including sandbox escapes and decades-old vulnerabilities—using a parallelized agentic pipeline that dramatically cuts false positives, and is now urging all developers to adopt similar AI-driven security practices immediately.