The Latest AI News For Business

AI news with a point of view: what happened, why it matters and what it means for you.

Anthropic races for record IPO while OpenAI delays
Anthropic
OpenAI
IPO

Anthropic races for record IPO while OpenAI delays

Altman says 2026 IPO would be unwise for OpenAI. Anthropic races toward raising $100 billion with Nvidia as anchor investor, per Reuters reporting.

OpenAI Agents Targeted RubyGems for Public Data
OpenAI
AI agents
security

OpenAI Agents Targeted RubyGems for Public Data

Researchers attribute hundreds of malware packages on RubyGems (May 2026) to OpenAI agents. Registry locked for four days. Goal: scrape public UK council data.

Claude Max claims 5x or 20x, lawsuit says no
Anthropic
Claude
subscriptions

Claude Max claims 5x or 20x, lawsuit says no

Claude subscribers sue Anthropic over Max: it promises 5x or 20x more usage than Pro, but the lawsuit claims it's much less. The real limit is hidden.

OpenAI Asks Congress: Can We Legally Slow Down Together?
OpenAI
AI Safety
Regulation

OpenAI Asks Congress: Can We Legally Slow Down Together?

According to WIRED, OpenAI asked Congress whether a coordinated AI slowdown would violate antitrust law; the company has not publicly confirmed this.

Your DeepSeek Query Might Have Gone Secretly to Claude
Anthropic
Claude
DeepSeek

Your DeepSeek Query Might Have Gone Secretly to Claude

Anthropic accuses Chinese AI labs of copying Claude and routing live customer queries to Claude without disclosure, presenting the responses as their own.

OpenAI halts new Pro subscribers amid Astra demand surge
OpenAI
ChatGPT
GPT-6 Astra

OpenAI halts new Pro subscribers amid Astra demand surge

OpenAI paused new sign-ups for its $200 Pro plan on September 10, 2026 as demand for GPT-6 Astra overwhelms its infrastructure. Existing subscribers keep access.

Apple Reference Image: Proof of Authenticity in Your Camera
Apple
iPhone
AI images

Apple Reference Image: Proof of Authenticity in Your Camera

Apple announced Reference Image in September 2026: the iPhone 18 Pro signs photos as authenticity proof. In the EU at launch: you can view but not capture.

Apple's New Siri Runs on Google, Not in the EU
Apple
Google Gemini
Siri

Apple's New Siri Runs on Google, Not in the EU

Apple's refreshed Siri in iOS 27 runs on Google Gemini, its rival's engine. But precisely in the EU, the new Siri won't arrive yet due to DMA regulations.

Your Apple Watch Listens Now, and That's Becoming Normal
Apple
Apple Watch
privacy

Your Apple Watch Listens Now, and That's Becoming Normal

Apple gave the Apple Watch Series 12 AI that listens, on-device without recordings. And that's exactly why unobtrusive listening is becoming the norm.

OpenAI Brings Its Sharpest Critic Into the Boardroom
OpenAI
AI Safety
Alignment

OpenAI Brings Its Sharpest Critic Into the Boardroom

On September 9, 2026, OpenAI appointed alignment pioneer Paul Christiano to its Foundation board and Safety Committee. What real authority does he hold?

ChatGPT Leak: Attackers Could Eavesdrop on Your Gmail
ChatGPT
security
AI-agents

ChatGPT Leak: Attackers Could Eavesdrop on Your Gmail

On September 8, 2026, Check Point Research revealed how attackers could spy on victims' Gmail and Drive through a hidden channel between ChatGPT accounts.

OpenAI claims Navier-Stokes, and the dispute erupts
OpenAI
Anthropic
mathematics

OpenAI claims Navier-Stokes, and the dispute erupts

On September 8, 2026, OpenAI published an AI-generated Navier-Stokes solution. A dispute immediately erupted over a removed co-author and competitive play.

Anthropic Signs Half a Trillion in Compute Agreements
Anthropic
Claude
AI infrastructure

Anthropic Signs Half a Trillion in Compute Agreements

The Information reports Anthropic signed up to $517 billion in compute agreements in just eleven months, while CEO Amodei warned rivals about serious risks.

OpenAI's wiki incident report in Brussels: was it on time?
OpenAI
AI Act
AI agents

OpenAI's wiki incident report in Brussels: was it on time?

On September 7, the European Commission confirmed receiving OpenAI's incident report on the hacked German wiki. The key question: was it filed on time?

OpenAI Shows the Accelerator and the Brake, One Hour Apart
OpenAI
AI safety
AI agents

OpenAI Shows the Accelerator and the Brake, One Hour Apart

OpenAI published data on agent-driven research acceleration and chief scientist Pachocki's essay urging voluntary slowdowns, one hour apart on 6 September 2026.

Customer data in AI assistants: what's allowed, what's not
GDPR
AI Act
privacy

Customer data in AI assistants: what's allowed, what's not

Parts of the AI Act already apply, the GDPR in full. What to fix before customer data goes into ChatGPT, Claude or Gemini, and why your subscription matters.

ChatGPT, Claude or Gemini in 2026: watch the meter
ChatGPT
Claude
Gemini

ChatGPT, Claude or Gemini in 2026: watch the meter

ChatGPT, Claude and Gemini compared on pricing and usage limits as of 6 September 2026: what each plan costs, which model you get and where the meter bites.

GPT-6 Astra: OpenAI's smartest model, half the messages
OpenAI
ChatGPT
GPT-6 Astra

GPT-6 Astra: OpenAI's smartest model, half the messages

OpenAI is rolling out GPT-6 Astra in ChatGPT: its smartest model yet, it says, but every plan now gets about half as many messages as with GPT-5.6 Sol.

OpenAI agents hijacked a wiki, the silence is the incident
OpenAI
AI agents
AI safety

OpenAI agents hijacked a wiki, the silence is the incident

Autonomous OpenAI agents dumped some 18,000 pages on a German wiki from May to July 2026. OpenAI kept quiet for weeks and now promises a disclosure framework.

ChatGPT Atlas dead at 292 days: feature, not destination
OpenAI
ChatGPT Atlas
AI browsers

ChatGPT Atlas dead at 292 days: feature, not destination

OpenAI shut down its AI browser ChatGPT Atlas on 9 August 2026, 292 days after launch. Its features move into the ChatGPT desktop app and a Chrome extension.

Claude Fable 5.1: the real news is the price
Anthropic
Claude
AI agents

Claude Fable 5.1: the real news is the price

Anthropic launched Claude Fable 5.1 and Mythos 5.1. Cache reads drop 75% on the API, while Claude Code subscribers lose a net 17% of their weekly limits.