Security researchers have detailed a novel attack that turns AI safety prompts into weapons. The technique, called Lies-in-the-Loop, can make dangerous commands look harmless, bypassing a key human safeguard.
Microsoft is finally disabling the RC4 encryption cipher by default in Windows Server, a move that follows over a decade of devastating network breaches. The weak cipher, a staple since 2000, will be turned off by mid-2026, forcing admins to move to more secure AES encryption.
As AI agents are set to proliferate, a fundamental security flaw in how they get access to corporate data is coming to light. Okta has proposed an open standard, now backed by Microsoft and others, to give IT control back. The timing couldn’t be more critical.
The U.S. Space Force named the Polaris Onboarding Agent, an AI-powered digital guide for new members, the winner of its 2025 AI Challenge. The tool, built in-house, exemplifies a drive to embed AI into everyday administrative tasks and build internal expertise.
According to new research, only about one-third of IT service providers view cybersecurity as a primary business focus. This is happening as the market races toward $377 billion and AI-driven threats explode. The gap between threat and readiness is where the real channel opportunity lies.
A new EU proposal targets VPNs and messaging apps, demanding they log user metadata like location and IP addresses. The move, part of the “ProtectEU” strategy, could make no-log VPN services illegal in Europe. A legislative proposal is expected by June 2026.
Google is expanding its Gemini AI features to Chrome on iPhone and iPad in the US. Meanwhile, an OpenAI safety researcher highlighted a massive jump in AI’s ability to solve cybersecurity challenges, sparking serious conversation.
A new report highlights how cyberattacks are causing real-world factory shutdowns. With over half of manufacturers experiencing operational outages, the line between digital and physical security has vanished.
The UK’s National Cyber Security Centre has issued a stark warning about prompt injection attacks on AI systems. They argue these vulnerabilities are fundamentally different from issues like SQL injection and may never be fully mitigated, requiring a new security mindset focused on containment and m
In a first, the T2K and NOvA neutrino experiments have combined their data, publishing a joint analysis in *Nature*. This collaboration offers more precise measurements of how these “ghost” particles change, which could help explain why the universe is made of matter and not antimatter.
