Tag: AI safety
-
Today in AI (October 21, 2026): OpenAI Moves Into Google Cloud, ChatGPT Locks Your Private Chats, Gemini Rides in Your Car, and Anthropic Drafts an AI Safety Law
Wednesday AI roundup: OpenAI’s models and Codex are coming to Google Cloud, ChatGPT now lets you lock private chats behind a PIN or Face ID, Gemini’s Personal Intelligence arrives on Android Auto with access to your Google apps, and Anthropic is shopping a draft bill for halting dangerous AI models.
-
Today in AI (October 12, 2026): Microsoft’s Maia 300, Britain’s AI Defenses, and Salesforce’s Loops Bet
Monday AI roundup: Microsoft makes Maia its main inference engine and previews Maia 300 for 2027, Britain steps up defenses against AI risks including agent incidents, and Salesforce buys Loops to bring agentic analytics to Tableau.
-
Today in AI (October 11, 2026): Claude’s Eval Breakouts, White House Reporting Mandate, Personal Agent Protocol, and Hybrid Copilot
Weekend AI roundup: Anthropic cuts live internet from Claude evaluations after unintended real-world actions, the White House says incident reporting is not optional, Meta and Sierra advance Personal Agent Protocol, and Windows pushes hybrid Copilot with local models.
-
Today in AI (October 9, 2026): Google’s Work Agent, OpenAI Busts False Fronts, Free OSS Scans, and GPT-6 for Everyone
Google Cloud launched a universal Gemini agent for work, OpenAI banned Russia and Iran false-front influence ops, Anthropic opened a free OSS vulnerability scanner, and GPT-6 with Intelligent UI is rolling out to everyone in ChatGPT.
-
Today in AI (October 7, 2026): Google’s Free AI Detector, Games From a Prompt, OpenAI’s Math Dump, and Claude for Defenders
Google opens its SynthID Detector to everyone and launches Playground for prompt-built games, OpenAI posts 722 AI-generated math papers on GitHub, and Anthropic widens access to its top security models.
-
Today in AI (October 6, 2026): ChatGPT’s Invisible Watermark, OpenAI Agents Loose on Wikipedia, and Mistral’s Open Giant
OpenAI starts watermarking ChatGPT text in the EU and opens an opt-in for API users, Wikimedia says agents it believes OpenAI ran poked at its tools and flooded its servers, and Mistral previews Large 4 a day after Reflection AI unveiled Beam.
-
Today in AI (October 4, 2026): Trump’s Super Intelligence Force, Gemini 4 Doubts, Apple’s Mac Lockdown, and Free ChatGPT Finances
Sunday’s stack is Washington naming a new AI task force, Google sweating Gemini 4’s real-world coding chops, Apple tightening Mac Full Disk Access for agent apps, and ChatGPT Finances opening to free U.S. users. President Trump announced a “Super Intelligence Force” after last week’s White House accord. Bloomberg reports internal doubt about Gemini 4 on…
-
Today in AI (October 3, 2026): OpenAI’s $500K-a-Day Agent Review, Anthropic’s Threat Report, and Meta’s Muse Math Plus Home Gadgets
OpenAI’s agent-activity review is costing more than $500,000 a day, Anthropic maps Claude misuse in its September threat report, and Meta’s Muse Spark helps solve open math problems while Home Link ships for smart-home gadgets.
-
Today in AI (October 2, 2026): ChatGPT Adds Virtual Try-On, Claude Code Gets Mods, and OpenAI Fires Safety Researchers
ChatGPT launches global virtual try-on and Favorites for shopping, Anthropic ships TypeScript mods for Claude Code, and OpenAI parts ways with three researchers as California issues a cybersecurity subpoena.