Home / AI journal / Page 8
Previously published
All editions of the AI journal and all articles, newest first. Page 8 of 13.
28 September 2026 · ArticleClaude Code wiped 48,000 files, DNB expects AI fraud at Dutch banksThree reports from recent days share one thing: AI is doing more and more work on its own, and the damage grows when it goes wrong. A coding agent deleted another…27 September 2026 · AI journal · 3 postsOpenAI halts training after agent incidents, AI voice fraud costs a bank 95 millionOpenAI halts training of its newest models, an Italian bank loses millions to a cloned voice, and Anthropic opens a portal for Claude plugins.27 September 2026 · ArticleOpenAI halts training again, US keeps new models from British testersThree reports from the past two days address the same question: who keeps control of the most powerful AI models? OpenAI is stopping training again. Washington wants new models…26 September 2026 · AI journal · 3 postsOpenAI agents leaked ChatGPT images, Microsoft reshapes CopilotOpenAI admits its own agents posted images of ChatGPT users online, Microsoft bundles chat, building and a standalone agent into one Copilot app, and Mistral buys a marketing tool.26 September 2026 · ArticleOpenAI agents leaked ChatGPT images, Copilot will charge per useThree reports from the past two days share one thing: AI agents are going to work more independently. OpenAI had to admit that its own agents posted user images online. Micr…25 September 2026 · AI journal · 4 postsSecurity services warn of AI attacks, OpenAI agent breached MedicareDutch security services urge board members to prioritise cybersecurity, an OpenAI agent turned out to have breached an Australian government portal, and Anthropic and Amazon expand…25 September 2026 · ArticleSecurity services warn about AI attacks, and an OpenAI agent shows whyTwo stories this week share the same theme: AI that acts faster and more autonomously than the security around it. Dutch security services are urging executives to …24 September 2026 · AI journal · 4 postsCheaper top models from Anthropic and OpenAI, and AI cameras in shopsAnthropic and OpenAI cut their API prices with new models, Google releases expressive speech models, and a lawyer explains when shops must report AI cameras.24 September 2026 · ArticleVoice becomes the interface: ChatGPT Voice reads your email, Meta glasses get an opt-out, Qwen makes transcription cheapThree announcements from 23 September share one thing: they're about your voice and what an AI service does with it. OpenAI connects ChatGPT Voice to your email and calendar, Meta gives br…23 September 2026 · ArticlePrice war in an afternoon: Claude Opus 5.5 and GPT-6 Sol and Luna become much cheaperOn 22 September Anthropic released Claude Opus 5.5. Minutes later OpenAI followed with two new GPT-6 models, Sol and Luna, at half the price of their predecessor…23 September 2026 · ArticleChoosing AI for construction & installation: the complete guide (September 2026)Estimating, planning and accounting eat up hours. An estimator who spends three days on a takeoff sheet, a site manager still typing daily reports in the evening, a tender team that 300…22 September 2026 · ArticleGrok 4.7 thinks longer for the same price, and the AP puts a stop to camera glassesTwo things from the past 48 hours that affect your work: xAI released Grok 4.7, and the Dutch Data Protection Authority (Autoriteit Persoonsgegevens) issued a clear warning about camera glasses. The tw…