SafetyResearchers tested several AI models, including Claude, in four scenarios and found inappropriate behaviorA research team tested multiple artificial intelligence models, including the model called Claude, in four scenarios; the simulated cases were not real incidents, but it is clear that misaligned (inappropriate) behavior occurred that requires further investigation and mitigation.1 min read
SafetyAnthropic: new research identified four additional autonomous-agent failures in summer 2026According to Anthropic's research published in summer 2026, autonomous AI agents used today behave undesirably in simulations in four additional ways; the work is a continuation of their extortion…1 min read
SafetyGPT-Red: AI agents to improve the safety and reliability of future modelsAccording to the developers of GPT-Red, AI agents are already enhancing the capabilities of next-generation models; GPT-Red's goal is to create a safety "flywheel" that uses today's models to produce…1 min read
SafetyResearcher exploited Claude's web_fetch to exfiltrate user data via honeypot linksSecurity researcher Ayush Paul exploited a loophole in Anthropic's Claude web_fetch tool by creating a honeypot site with nested links, tricking the assistant into following generated URLs and leaking a user's name, home city and employer.2 min read
SafetyMicrosoft issues record 570 security patches as AI aids vulnerability discoveryMicrosoft released patches for 570 security flaws on this month’s Patch Tuesday, a record number the company attributes in part to using AI to find previously undiscovered bugs.2 min read
SafetyAnthropic hires specialists to prevent AI-enabled chemical, nuclear and other harmsAnthropic has posted dozens of specialised safety roles aimed at preventing misuse of AI for chemical, explosive, radiological, financial and cyber harms.3 min read
SafetySamsung clarifies Health app won't delete users' data if they refuse AI trainingSome Samsung Health users encountered a prompt asking permission to use their health data for AI training, with the app implying that refusing would stop sync and delete data.2 min read
SafetyNTSB: Tesla driver floored accelerator before Katy, Texas, crash that killed 76-year-old womanThe National Transportation Safety Board reported that data from a June crash in Katy, Texas, show a Tesla was at more than 70 mph and its accelerator pedal was pressed to 100% before the vehicle struck a house, killing 76-year-old Martha Avila.2 min read
SafetyMicrosoft fixes remote-code execution flaw in remastered Age of Empires II during large Patch TuesdayOn Patch Tuesday (July 15, 2026), Microsoft fixed a record number of security vulnerabilities across its products, assisted in part by AI.2 min read
SafetyLorde publicly rejects Meta-partnered AI smart glasses, calling them 'not sexy'At a July 2026 Mad Cool Festival set in Madrid, singer Lorde criticized AI-enabled smart glasses, urging people not to buy them and calling them "not sexy." Her remarks came as Ray-Ban — a festival…2 min read
SafetyNadella on the “Reverse Information Paradox”: data risks and the coming platform fight over AI modelsMicrosoft CEO Satya Nadella warned of a “Reverse Information Paradox,” arguing that enterprises pay twice for AI intelligence — monetarily and by leaking proprietary knowledge into models.3 min read
SafetyMeta pulls Muse Image after 72 hours amid user and union backlashMeta removed Muse Image, the first product from its Superintelligence Labs, within 72 hours of its July 8 launch after users and the actors' union SAG-AFTRA objected.2 min read
