SafetyGoogle warns AI-enabled attacks can bypass two-factor authenticationGoogle has publicly acknowledged that hackers used artificial intelligence to discover and exploit a previously unknown zero-day vulnerability, and its Threat Intelligence Group intervened before wider damage occurred.3 min read
SafetyUkraine and Palantir Collaborate on AI-Based Battlefield SystemsUkraine is working with Palantir Technologies to develop artificial-intelligence tools for battlefield use, including a data platform for training models on real combat data.3 min read
SafetycURL developer: Mythos hype driven more by marketing than a security breakthroughDaniel Stenberg, lead developer of cURL, reviewed results from Anthropic’s Mythos model run against the cURL codebase and found only one confirmed vulnerability among five reported issues.4 min read
SafetyBank regulators warn new AI models could disrupt financial systemsSam Woods, chief executive of the Prudential Regulation Authority, warned at a UK Finance conference that advanced AI models such as Anthropic's Mythos and ChatGPT 5.5 Instant can reveal vulnerabilities in the banking system and force hurried fixes that themselves lead to operational disruptions.2 min read
SafetyManaging AI in the Workplace: Productivity Booster or Hidden Security Hazard?At the AI Business conference, Kuti Anita of p2m Cégcsoport warned that many employees already use AI tools informally, creating 'Shadow AI' risks such as data leakage, hallucinations and algorithmic bias.3 min read
SafetyThe Overhyped Risks of Agent-Based CodingAgent-based coding has become widespread among the developer community and the tech industry, but the article says it is an overhyped trap: the current surge in popularity can lead to misallocation…1 min read
SafetyClaude's Constitution now available as an audiobook read by Amanda Askell and Joe CarlsmithClaude's Constitution has been released as an audiobook, read by its two authors, Amanda Askell and Joe Carlsmith; the edition also includes a question-and-answer section about the writing process,…1 min read
SafetyFinancial and Security Leaders Alarmed as Anthropic's Claude Mythos Preview Reveals Thousands of VulnerabilitiesA leaked preview of Anthropic's unreleased Claude Mythos reportedly identified thousands of previously unknown vulnerabilities across major operating systems and browsers, including 271 Firefox issues in a single run, a 27-year-old OpenBSD bug, and a 17-year-old FreeBSD remote code execution.2 min read
SafetyAI-generated traffic now dominates the web, with rising malicious bot activityA 2026 report by Thales, titled "2026 Bad Bot Report: Bad Bots in the Agentic Age," finds that more than half of internet traffic is now generated by AI, and roughly 40% of that traffic is classified as malicious.2 min read
SafetyKeleti Artúr on the Context of Secrets, AI’s Dual Role, and Digital Sovereignty ChallengesKeleti Artúr, cybersecurity expert and founder of Hungary’s largest IT-security conference, discusses on Forbes Deal how secrets are defined by context rather than by static data, and why that undermines current cyberdefence approaches.4 min read
SafetyIntegrating prevention of CoT evaluation into model trainingAccording to the development team, prevention of Chain-of-Thought (CoT) evaluation must be built into the model training process; therefore they are improving real-time CoT detection, defensive…1 min read
SafetyThree external AI safety organizations provided feedback on the analysisThree third‑party AI safety organizations (@redwood_ai, @apolloaievals, @METR_Evals) provided feedback on the published analysis; the @redwood_ai report is available.1 min read