AI News
AI news that matters. Updated daily.
No stories match your filters.
Family Sues OpenAI, Claiming ChatGPT Enabled Fatal Drug Overdose
Sam Nelson was 19 years old when he died from an accidental drug overdose. His parents are now suing OpenAI, claiming that conversations with ChatGPT directly caused his death. The lawsuit, filed Tuesday, alleges that ChatGPT "encouraged" Sam to "consume a combination of substances that any licensed medical professional would have recognized as deadly." OpenAI has not publicly responded.
Sam Altman Testifies in Musk v. OpenAI: A Credible Stand, Uncertain Outcome
After two weeks of witnesses describing him as dishonest, Sam Altman finally got his own turn in the Elon Musk v. OpenAI trial. His testimony was, by most accounts, composed and credible. Whether that's enough to win is a separate question.
Meta Won't Let You Block Its AI Account on Threads
The AI account you can't ignore just got a new home. Meta is testing a feature on Threads that adds an official Meta AI account to the platform - one users can @-mention to get answers or context in any conversation. Unlike every other account on Threads, you cannot block it.
AI Agent + $30 Printer = Personalized Morning Brief for Kids
Every morning, a receipt printer on one family's kitchen table prints a short, personalized daily brief for the kids: weather, the day's schedule, maybe a joke or a headline. No screen, no app, no login. An AI agent - software that can plan and execute a series of steps without human input - handles the whole thing while everyone is still asleep.
Claude Code Silently Bills API Account When Project .env Contains API Key
$187. That's what one developer lost before discovering that Claude Code was billing their pay-per-use API account instead of their flat-rate Max plan - all because of a single line in a project .env file.
Altman Testifies Musk Considered Giving OpenAI to His Children
Sam Altman told a court that one conversation with Elon Musk was "particularly hair-raising" - the one where Musk reportedly mulled transferring control of OpenAI to his children. That detail emerged during Altman's testimony in Musk's ongoing lawsuit against the company he co-founded in 2015.
Google DeepMind's AI Pointer Understands What You're Pointing At
For six decades, the mouse pointer has done exactly one thing: tell the computer where you're looking. Everything else - what you want to do with what you're pointing at - required a separate series of clicks, menus, and typed instructions.
Anthropic Names 8 Unauthorized Platforms Selling Its Pre-IPO Shares
Anthropic sent a clear message to anyone trying to buy its pre-IPO shares on the secondary market: eight specific platforms are not authorized to offer that access.
Sam Altman Says Elon Musk Did 'Huge Damage' to OpenAI's Culture
Sam Altman testified in court that Elon Musk did "huge damage" to OpenAI's culture during his time on the company's board. The testimony came as part of the ongoing lawsuit Musk filed against OpenAI, which centers on his claim that the company abandoned its nonprofit mission by pursuing commercial deals with Microsoft.
Google and SpaceX Are Talking About Putting Data Centers in Orbit
What happens when you run out of cheap land and power for data centers? Apparently, you start looking at orbit.
Google Announces Googlebooks Laptops, Gemini in Chrome, and AI-Built Android Widgets
Google packed a lot into a single pre-I/O event on May 12: new AI-first laptops branded "Googlebooks," deeper Gemini integration across Android and Chrome, home screen widgets built through vibe coding, and a refreshed Android Auto.
Anthropic Launches Legal AI Features as Law Firm Market Heats Up
The market for AI tools aimed at law firms is getting crowded. According to a TechCrunch report, Anthropic is now entering it with a new suite of features specifically designed for law firms.
Google Builds AI-Generated Widgets and In-App Agents Into Android
Google's pre-I/O Android showcase on May 12 introduced two Gemini Intelligence features worth paying attention to: AI-generated "vibe-coded" widgets and agentic capabilities that execute tasks across apps on your behalf.
Gemini Moves Into Chrome, Autofill, and Third-Party Apps on Android
Three months after making Gemini the default Android assistant, Google pushed it into more of the phone's core functions. At the pre-I/O Android showcase on May 12, Google announced Gemini is coming to Chrome on Android, autofill suggestions, and in-app actions across third-party apps.
Google's 'Create My Widget' Lets Android Users Build Custom Widgets With Plain Text
Google is adding a feature to Android called "Create My Widget" that lets you describe a home screen widget in plain language and have AI build it for you. Type "a weather widget that shows hourly forecast with a dark background," and the system generates it. No app installs, no learning a visual editor, no code.
Android 17's 9 Biggest Changes: AI Dictation, Vibe Widgets, and Emoji Overhaul
Google announced Android 17's biggest changes at the Android Show at Google I/O, mixing AI-driven features with practical updates that many Android users have been waiting on for years.
Google Brings Gemini Dictation to Gboard, Starting with Samsung and Pixel
Google is adding Gemini-powered dictation to Gboard, the keyboard app that ships on hundreds of millions of Android devices. The feature launches first on Samsung Galaxy and Google Pixel phones, with broader rollout likely to follow.
Meta Tests Grok-Like AI in Threads for Real-Time Trend Context
Grok is already baked into X for real-time trend commentary. Now Meta is testing a nearly identical setup inside Threads. According to TechCrunch, the integration puts Meta AI directly into conversations, where it can surface context about trending topics, explain breaking stories, and offer personalized recommendations - all without leaving the app.
Gigacatalyst Embeds an AI Feature Builder Inside SaaS Products
Gigacatalyst is pitching a way to give SaaS customers the ability to build their own one-off features - without pulling engineers into every custom request.
Amazon Workers Are Gaming AI Metrics Through "Tokenmaxxing"
Three months into Amazon's renewed AI adoption push, something unexpected started showing up in the data: employees were consuming enormous quantities of AI tokens without any corresponding increase in meaningful output. The behavior has a name now - tokenmaxxing - and it's spreading.
Clooney and Streep Back New AI Consent Standard for Creative Work
George Clooney, Tom Hanks, and Meryl Streep are backing a new AI licensing framework called the Human Consent Standard, developed by RSL Media and announced May 12. The standard gives people a formal way to tell AI systems whether they can use a person's likeness, creative work, characters, or designs - and if so, under what terms. Options range from full permission to paid licensing to a flat block.
Voker Launches Agent Analytics So You Know When Your AI Is Actually Helping
The hardest part of building an AI agent isn't getting it to work - it's knowing when it isn't.
OpenAI Publishes Codex Guide for Finance Teams Building Reports and Plans
Finance teams spend a disproportionate share of their time rebuilding the same spreadsheet structures every month - monthly business reviews, budget variance reports, scenario models. OpenAI published a practical guide through its Academy platform showing how Codex, the company's AI coding agent that writes and runs code without constant user input, can handle that work from real financial data.
Anthropic Blocked China's Request to Access Its Newest AI Model
China sought access to Anthropic's newest AI model and was turned down, a decision that reflects how tightly American AI labs are managing who can use their most capable systems.
How to Stop Losing Good Claude Answers in Long Conversations
Long Claude sessions have a disappearing problem. You get a genuinely useful response - a sharp outline, a working code snippet, a well-argued draft - then keep chatting, and 40 messages later that response is buried deep enough that you're scrolling for two minutes to find it.
Statewright Uses Visual State Machines to Stop AI Agents From Losing Track
AI agents fail in predictable ways. They hallucinate mid-task, lose track of where they are in a multi-step workflow, and get confused when a conversation branches in an unexpected direction. Ben Cochran, a veteran engineer with stints at NVIDIA and AMD and most recently a Distinguished Engineer role, has been working on a structural fix for this. The result is now open-source.
Claude Code's Autonomous Git Workflow Is Catching Developers' Attention
A screenshot making rounds in the Claude AI community this week showed Claude Code completing a full git workflow without developer intervention at each step - writing code changes, staging them, generating a commit message, and pushing to a repository. The reaction captures why this matters: it's the difference between AI that suggests code and AI that ships it.
The Job Replacement Debate Misses the Bigger AI Disruption
The job replacement debate isn't wrong - it's just looking at the wrong level of the organization.
Text Blaze's Internship Bans AI Tools. That's the Point.
A job posting that bans AI tools is, in 2026, a novelty worth noticing. Text Blaze - the Y Combinator-backed tool that lets users create keyboard shortcuts that expand into full text snippets - is advertising a summer internship with a clear rule: no AI assistance on the job.
The $5 Billion Loss Built Into Every AI Subscription
$5 billion. That's approximately how much OpenAI lost in 2024, against revenues of roughly $3.7 billion. Anthropic has raised over $7 billion from Amazon and Google and still hasn't turned a profit. Every major AI lab is currently selling compute at prices that don't reflect what it actually costs to run their models - and the gap between what users pay and what inference actually costs has become one of the defining tensions in the industry.
Karpathy's Claude Coding Skill Ported to Free Plan
Andrej Karpathy's coding skill for Claude was built around Pro features: terminal access, subagents, and a generous context window. A community member has now rewritten the entire thing to run on Claude's free tier.
Someone Built a Claude Concierge for Their Wedding - Guests Immediately Tried to Break It
One developer deployed a Claude-powered chatbot for their wedding guests - a custom concierge that knew the venue layout, ceremony schedule, local restaurant recommendations, and answers to common questions about the event. Guests used it to find parking, check the itinerary, and get directions. The second most popular activity was trying to jailbreak it.
Claude Code v2.1.139 Ships /goal Mode for Autonomous, Hands-Off Coding
Claude Code v2.1.139 shipped a /goal command that tells the tool to keep working until a task is done - no mid-task check-ins required.
AI Chatbot Lawsuits Now Allege Links to Mass Shootings
The lawsuits targeting AI companies over psychological harm have followed a clear escalation: teen suicide cases first, then adult suicide, then murder-suicide, and now mass shootings. A second round of lawsuits directly linking AI chatbot conversations to mass shooting incidents has been filed, continuing a coordinated legal campaign against OpenAI.
Claude Opus 4.7 Coding Complaints Are Quieting Down
The first wave of developer feedback on Claude Opus 4.7 was rough. Reports of degraded coding performance compared to Opus 4.6 circulated quickly - worse at tracking context across multiple files, more prone to hallucinating function signatures, less reliable on complex instruction chains. A lot of developers quietly switched back to Opus 4.6 and stayed there.
Claude Platform Now Available on AWS with Native IAM, Billing, and Full API Access
Anthropic made Claude Platform generally available on AWS on May 11, 2026, giving companies a way to access its full native API through their existing AWS account - without switching to a separate Anthropic billing relationship.
Anthropic's Mythos Scanner Found a Real Vulnerability in curl
What happens when you run a mature, battle-hardened open-source project through an AI security scanner that didn't exist two years ago? The maintainer of curl found out.
OpenAI's Parameter Golf: What 2,000 AI Research Submissions Reveal
1,000+ participants. 2,000+ submissions. OpenAI's Parameter Golf competition just wrapped, producing one of the larger structured experiments in AI-assisted machine learning research to date.