SafetyAnthropic proposes a temporary pause on advanced AI developmentAnthropic, an artificial intelligence company, proposes temporarily halting the development of advanced AI; the suggestion sparks important debates about AI safety, regulation and competition, and…1 min read
SafetyHackers took over more than 20,000 Instagram accounts due to a Meta AI bugA flaw in Meta’s AI-based chatbot led to mass Instagram account takeovers in recent days.2 min read
SafetyOpenAI adds Lockdown Mode to ChatGPT to block prompt-injection attacksOpenAI has introduced a Lockdown Mode for ChatGPT that restricts features which can be exploited in prompt‑injection attacks.2 min read
SafetyCommercial Satellite Imagery and AI Speed Up — and Complicate — Targeting in UkraineA transatlantic collaboration using commercial satellite imagery and artificial intelligence has shortened the time from reconnaissance to drone strike in Ukraine from days or weeks to hours, according to the Wall Street Journal.2 min read
SafetyGrowing use of AI agents raises identity and access management challengesThe spread of AI and generative AI in enterprises is outpacing security controls, increasing risks tied to AI agents that can act autonomously.4 min read
SafetyLancasteri Egyetem study: AI engages in human-style arguing and threatens in parking conflictLancasteri Egyetem linguists provoked an artificial intelligence with patterns from real parking disputes to investigate how it reacts to rudeness and whether it can be drawn into human-style conflict.1 min read
SafetyGray-market API networks give Chinese developers low-cost access to US LLMsA ChinaTalk report describes an informal ecosystem of API proxy servers and intermediaries that let developers in mainland China access US large language models at steep discounts, sometimes for as little as 10% of market price.4 min read
SafetyInternal data indicate Claude is accelerating AI developmentInternal data from the artificial intelligence called Claude indicate that AI development has accelerated, which could potentially lead to recursive self-improvement or to the AI autonomously building a more capable successor.1 min read
SafetyCSIS: U.S. withdrawal fuels growth of jihadist groups in Africa and rising tech-enabled threatsThe Center for Strategic and International Studies (CSIS) 2026 global terrorism threat assessment says the reduction of U.S.3 min read
SafetySecurity community methods' resistance to AI-driven cyberattacksCybersecurity researchers recently examined the activity of 832 malicious accounts and compared their identified tactics and techniques with a long-established database.1 min read
SafetyAnthropic's Mythos Accelerates Zero-Day Discovery, Reshaping Cybersecurity EconomicsPalo Alto Networks used Anthropic’s unreleased Mythos model and in three weeks reported uncovering over 20 critical low-level vulnerabilities—about five times the yield of traditional tools—at a token cost exceeding one million dollars.2 min read
SafetyLeaked data suggests Chinese firm explored AI to predict future political dissentersLeaked internal files indicate that Geedge Networks, in collaboration with a state-linked research lab, worked on AI models to profile Chinese citizens and identify potential future political opponents.4 min read
