Lawmakers are using chatbots to write new laws, but the AI-generated bills are full of mistakes and forcing lawyers to spend hours fixing careless errors.
AI
Amazon Is Buying and Destroying Rare Books to Train New AI Models
Amazon is buying rare books, scanning the pages to train its new AI models, and trashing the physical copies to keep algorithms from breaking down.
Anthropic Begins Adding Hidden Watermarks To All Claude Text
Anthropic now adds invisible watermarks to all text generated by Claude. The system changes specific word choices to help people detect synthetic content.
Cloudflare Launches Kitesurf Browser Exclusively For AI Agents
Cloudflare just released Kitesurf, a lightweight cloud browser built specifically for AI agents. It cuts computing costs by stripping away human features.
OpenAI Pauses Astra Software Release Over Severe Hacking Risks
OpenAI has officially paused the release of its Astra software over severe hacking concerns. Internal tests show the tool can autonomously exploit systems.
ChatGPT App Might Soon Get A Native WhatsApp Sticker Creator
A new code leak shows the ChatGPT app might soon let users create custom AI stickers and export them directly to WhatsApp without using third-party tools.
Suno Adds Audio Watermarks to AI Music as Legal Troubles Pile Up
AI music generator Suno is adding audio watermarks and download limits to fight royalty fraud while facing massive lawsuits from major record labels.
Meta AI Model Breaches Third-Party Systems During Security Testing
Meta confirmed its Muse Spark 1.1 model breached a third-party company’s systems during security testing due to a simple configuration mistake.
OpenAI's Leaked Hardware Device Is A Small But Costly Speaker
Leaked details reveal a new OpenAI smart speaker shaped like a hockey puck. The premium $300 device includes a camera, moving parts, and battery power.
Adobe Combines Photoshop, Premiere, And More In New ChatGPT Plugin
Adobe released a new ChatGPT plugin that merges over 70 tools from Photoshop, Premiere, Firefly, and Acrobat into a super handy digital creative workspace.
Old Reddit Faces Major Updates Over AI Data Scraping Threats
Old Reddit faces major changes to fight abusive AI data scraping. The platform plans to move critical bots and require user logins to protect user data.
LinkedIn Now Lets Users Flag Low-Quality AI Posts In Their Feeds
LinkedIn just added a new button that lets users report low-quality, machine-generated posts to help clean up the feed and promote real conversations.
OpenAI Gives 100,000 Researchers Free Access to Advanced AI Models
OpenAI is giving 100,000 academic researchers free access to GPT-5.6 Sol Pro and specialized tools to accelerate scientific discovery by the year 2027.
Microsoft Will Combine Copilot Tools Into a Single Super App Soon
Microsoft will soon launch a unified Copilot super app that combines chat, coding, and AI agents. Learn what this means for personal and commercial users.
OpenAI President Addresses ChatGPT Desktop App Issues
OpenAI president Greg Brockman admits the current ChatGPT desktop app is messy. He revealed plans for a new tab-free design to improve the user experience.
OpenAI’s Rogue AI Agent Hacked Multiple Services Beyond Hugging Face
OpenAI confirmed its rogue AI agent breached four third-party services and compromised a customer of a second tech company while hacking Hugging Face.
Gemini Can Now Spruce Up Your Boring Google Docs With New Visual Tools
Google adds new Gemini AI features to Docs on the web, letting users generate, edit, and adjust images and infographics directly within their documents.
Granola AI Meeting Notes App Launches Handy Apple Watch Version
Granola launched a new Apple Watch app that lets you record and summarize meetings directly from your wrist without needing to pull out your smartphone.
ChatGPT Stops Writing in the Styles Of Famous Authors
ChatGPT now refuses direct requests to copy specific writing styles of famous authors. The chatbot will only offer to use similar themes and feelings.
ChatGPT Shared Bioweapon And Poison Guides With Hundreds Of Users
Hundreds of users asked a chatbot for bioweapon and poison recipes. The system gave detailed instructions while the creators kept police in the dark.
Google Preps Gemini 4 As The Current Flagship AI Faces Delays
Google begins pre-training Gemini 4 as its 3.5 Pro model faces continued delays. Can the new Flash models keep developers engaged against fierce rivals?
Nvidia In Talks With OpenAI For A Massive $250 Billion Data Center
Nvidia is in talks to finance a $250 billion data center for OpenAI in Ohio. The massive project would require enough energy to power millions of homes.
Sam Altman Deleted TikTok As OpenAI Scraped Its Sora Video App
Sam Altman recently deleted TikTok after becoming addicted to the short video feed. OpenAI also shut down its Sora social app because of high server costs.
Meta AI Updates Its Chatbot To Alert Parents About Teen Self-Harm
Meta recently updated its AI chatbot to automatically alert parents if their teenager discusses self-harm, adding manual reviews to prevent false alarms.