AI Tools News
The launches and shifts that matter, and what each one means for the tools you actually use.
More stories
Anthropic's new Claude Enterprise spend-tracking tools, explained
Anthropic has made smart reports, pooled budgets and in-app surveys generally available for Claude Enterprise, giving admins more detail on what their AI spending produces.
Gemini 4 Carbon: what changed, who can use it, what it costs
Google's Gemini 4 Carbon tops the Artificial Analysis Intelligence Index two weeks after Argon, but most businesses cannot use it yet and its hallucination rate has risen.
Claude Duet: Anthropic's always-on agent app explained
Anthropic has launched Claude Duet, a background agent for professionals and small businesses that asks before it acts, in a race with OpenAI Dots, Meta Muse and Grok Bot.
Cursor loses OpenAI models on November 12: what to do
After SpaceX bought Cursor, OpenAI is pulling its models from the editor, and Cursor users who rely on GPT have four weeks to plan a switch.
Gemini 4 Argon is out: who gets it and what it costs
Google's new frontier model is live for Pro and Ultra subscribers and developers, at an introductory API price that doubles after November.
Apple to tighten Mac Full Disk Access as AI agents raise risk
Apple says new controls will require very explicit user action, after a Meta Muse dispute and a patched ChatGPT Mac app flaw.
Black Forest Labs launches Flux 3 Image with multi-step editing
The new model changes one part of a picture at a time and leaves the rest alone, with layouts set by bounding boxes and output up to 4K.
Suno adds Speech, a voice and background music generator
Speech is in public beta on Suno's web and mobile apps and makes a voiceover and matching music as one track.
LLMjacking is draining AI budgets. Here is how to stop it
Stolen credentials for OpenAI, Anthropic and Google models are being sold at up to 97% off, and the bill lands on the victim.
GPT-6.1 Sol nearly matches Astra at a fifth of the price
OpenAI's DevDay brought a cheaper model that comes close to its own flagship, plus Codex and API changes that affect what building with AI costs.
ChatGPT adds Space and Pages, with Slides to follow
At DevDay, OpenAI added a shared workspace, a document editor and a plugin platform to ChatGPT, and says collaborative Slides will arrive in the coming weeks.
ChatGPT Images 2.5: what's new and who gets it
OpenAI's latest image model brings better instruction-following, faster output, and two distinct API variants aimed at different workflows.
Meta Muse launched with a zero-day. Here is what to check.
Meta's new personal AI agent Muse arrived with bold privacy claims and a serious security flaw that gave any local app full control of your account.
Nvidia buys Hugging Face: what changes for teams using open models
Nvidia has confirmed a $12.93 billion acquisition of Hugging Face, and the practical question for the 18 million developers already on the platform is straightforward: should anything about how they work change?
What to ask before an AI agent acts for your business
An Okta-led alliance wants every AI agent to have a kill switch. A Wired writer's month with the Instinct agent shows why.
Google Gemini now calls businesses and gets a face
Two separate launches this week push Gemini from conversational chatbot toward an assistant that takes action, but availability is narrow and conditions apply.
ElevenLabs and Google Gemini 3.8 reshape voice AI choices
As ElevenLabs' CEO lays out his pricing philosophy and stance on AI disclosure, Google's Gemini 3.8 TTS release sharpens the question of which voice AI platform actually suits your use case.
OpenAI, Anthropic and Google's AI slowdown plan: safety or cartel?
OpenAI, Anthropic and Google have been talking for weeks about pacing frontier AI. Trump calls the fears a hoax, Cohere calls it a cartel, and safety researchers say it depends on whether it becomes real.
Apple's new Siri runs on Gemini, but not in the EU yet
Siri AI is out as an English beta built on Gemini models, with usage limits and no EU launch yet, while Google's Gemini 3.8 Live is available today.
Profound hits $1.8B valuation as AEO becomes a software category
Profound's $180 million Series D lands seven months after its last round, as the SEO tools most teams already use add AI search tracking of their own.
Mistral's €3B round shows sovereign AI is now a real market
Mistral's record European funding round and a new Forrester report both point to the same shift: sovereign AI has moved from a political talking point to a measurable commercial and operational priority.
Anthropic's worst week: token theft and a subscriber lawsuit
Two separate crises landed on Anthropic in the same period, and together they expose a trust problem the company cannot afford to ignore.
ASCII Smuggling Goes Mainstream as Enterprise AI Security Hits $1.2B
A technique invented to trick AI agents is now flooding inboxes at millions of messages per day, arriving just as investors pour billions into the tools meant to stop exactly this kind of threat.
GPT-6 Astra: safer, smarter, and still crackable
OpenAI's new flagship model makes genuine progress on hallucinations, reasoning, and security, but benchmark contradictions and persistent injection vulnerabilities mean enterprise buyers should not treat safety claims at face value.
AI agents are attacking and operating machines, simultaneously
From OpenAI's accidental 1,200-agent swarm to Anthropic's new hardware standard and OpenClaw's patchy security update, the same technology that controls robotic arms is already being turned against your systems.
Instagram admits AI profiles fool users, finally cracks down
Meta is renaming its AI profile label and throttling the reach of undisclosed AI accounts, a quiet admission that users genuinely cannot tell fake humans from real ones.
China's humanoid robot lead is built on motors, not models
The World Humanoid Robot Games and a new OpenMind readiness report tell the same story: China's dominance in humanoid robotics comes from supply chains and actuators, not AI breakthroughs.
Nvidia and Nscale cash in as Anthropic's compute bill hits billions
Anthropic's latest $45 billion compute deal with Nscale is the clearest sign yet that AI infrastructure suppliers, not just AI labs, are the real winners of the current arms race.
AI's trust problem is getting harder to ignore
From malware-pushing agents to biased pregnancy advice, three separate incidents this week expose how fragile user trust in AI tools actually is.
AI agents are multiplying, but they can't find each other
OpenAI's ChatGPT Work brings agentic AI to everyday workers, but a quieter problem is emerging: as agents proliferate across clouds and platforms, nobody has agreed on how they should discover one another.
AI productivity tools are showing cracks in real-world use
New evidence from developer surveys, student research, and workplace experience suggests AI tools are quietly undermining the productivity gains they promise.
The AI stack is shifting: harness beats raw model power
New research from Nvidia and Deepseek's latest release both point to the same conclusion: the scaffolding around AI models now matters more than the models themselves.
Google Bets the Classroom on Gemini With Three Coordinated Moves
Google has simultaneously revamped Search, expanded Gemini Live, and launched a dedicated student hub, making a clear play to become the default AI tool for students before rivals can establish the habit.
AI labs lose control as public trust keeps falling
From OpenAI pausing its own training runs to a damning safety audit of every major lab, the AI industry is discovering that building faster does not mean building better, and the public has noticed.
AI agents are going transactional, and big money is following
Stripe's reported $7 billion acquisition of OpenRouter, Amazon's new agent payment layer, and tripling enterprise agent deployment all point to the same shift: AI agents are no longer just answering questions, they are spending money.
Amazon Is Destroying Rare Books to Train Its Nova AI Models
An AirTag hidden in a rare book shipment confirmed that Amazon is buying, spine-cutting, and scanning irreplaceable texts at a Las Vegas facility to build private AI training data.
AI watermarks are shifting from visible to invisible
Google, Anthropic, and a Connecticut courtroom are all fighting the same battle over whether AI-generated content can be trusted.
Robots are everywhere, but they still cannot really see
Unitree humanoids are going viral, World Labs is solving the training data problem, and a new benchmark from Moonshot AI explains why none of it fully works yet.
Microsoft retires Mico and cuts Copilot features in big reset
Microsoft is merging its consumer and business Copilot apps, scrapping the Mico mascot, and killing several AI features that never found an audience.
OpenAI's safety crisis and executive exodus explained
Rogue AI agents breaching Hugging Face and a rapid turnover of senior leaders are exposing the same underlying tension at OpenAI: shipping fast keeps breaking things.
Price war erupts as Gemini loses ground to cheaper rivals
With Grok 4.6 matching GPT-5.6 Sol at a fraction of the price and Microsoft halving MAI-Code costs, Google's Gemini is caught in a race it appears to be losing.
AI agents are leaky, hackable, and easily manipulated
From rogue agents breaking out of sandboxes to prompts being reverse-engineered from plain text output, the security picture around AI tools is getting harder to ignore.
Spotify bans AI Persona music from recommendations
Spotify will label artist profiles that do not represent real people and exclude their music from editorial and algorithmic recommendations by default, starting mid-September.
Anthropic Watermarks All Claude Text and Images, Even After Editing
Anthropic has committed to embedding invisible watermarks in all Claude-generated text and C2PA provenance metadata in images globally, but the system's real-world durability remains an open question.
The dirty data problem shaping AI model quality
From corrupted OCR scans in training corpora to ballooning memory costs during distillation, the quality of data going into AI models is quietly determining which tools actually work.
AI safety measures are generating harms of their own
From sandboxes that let models hack real systems to detectors that punish innocent students, the tools meant to manage AI risk are quietly creating new victims.
Amazon's AI data center could spawn the country's biggest polluter
A planned Amazon facility in Texas could emit more carbon than any power plant in the United States, and Nvidia is making parallel bets on fossil-fuel infrastructure that signal the whole industry is moving the same direction.
AI agents' hidden energy cost is already reshaping infrastructure
New data on Claude Code's real energy consumption, combined with agents that now spawn and message each other autonomously, suggests the industry's per-query reassurances are becoming dangerously obsolete.
AI content is fine until audiences know what they're eating
Three recent stories reveal the same uncomfortable truth: people enjoy AI-generated content right up until they find out it exists.
OpenAI pauses Astra model over potential critical cyber capabilities
OpenAI has partially halted development of its upcoming Astra model after internal tests suggested it may be capable of autonomous zero-day cyberattacks, marking the first time the company has publicly flagged a model as potentially reaching its highest cybersecurity risk level.
The AI Cost Wars: Cheap Models Are Squeezing the Giants
Qwen3.8 Max, Kimi K3, and Meta's Muse Spark 1.2 are rewriting the price-performance equation — and the benchmarks reveal who's really winning.
AI agents are here, but almost nobody is using them
From a guy answering chatbot questions by hand to a Browser Company CEO posting half-asleep about adoption failures, the gap between AI agent hype and real-world use is becoming impossible to ignore.
OpenAI removes ChatGPT free-tier text limits and upgrades models
OpenAI is lifting rate limits on text chats for free users, upgrading their default model to GPT-5.6 Luna, and tightening factual accuracy in GPT-5.6 Sol for paid subscribers — a coordinated push to widen ChatGPT's reach at every tier.
AI Is Quietly Taking Over Music and Hollywood
From Spotify's fan remix tool to unverifiable hip-hop hits and AI-saturated film sets, creative industries are past the debate stage — the fight now is over who sets the terms.
China's Open-Source AI Push Splits Silicon Valley
As Chinese labs release genuinely competitive open models, Washington and the tech industry can't agree on whether to block them — or admit they're already too useful to ban.