AI NEWSROOMBY CLASSWISE.IO

Tuesday, September 22, 2026

OpenAI creates math advisory group after model

Plus: Amazon blocks Meta’s Muse shopping agent; Anthropic weighs a new model as OpenAI gains ground

Welcome in—today feels like a quiet Tuesday of chess moves and new alliances, so pull up a chair.

RESEARCH

🧮 OpenAI creates math advisory group after model claims

OpenAI creates math advisory group after model claims
Image source: techcrunch.com

OpenAI announced a new independent advisory group hosted at the Institute for Advanced Study in Princeton, New Jersey. The group is meant to give mathematicians more input into OpenAI's math-oriented research.

The details:

  • The group is called the Advisory Group on Mathematics and Artificial Intelligence.
  • OpenAI claims that the same internal model has resolved over 100 additional open problems across most areas of mathematics.
  • Earlier this month, 25 Fields Medal-winning mathematicians signed an open letter arguing that AI labs are threatening their intellectual work.

First reported by techcrunch.com, 8h ago · techcrunch.com · neowin.net

AGENTS

🛑 Amazon blocks Meta’s Muse shopping agent

Amazon blocks Meta’s Muse shopping agent
Image source: techcrunch.com

Meta’s AI assistant Muse started getting an error message when users tried to buy goods on Amazon. The error message stated that continued access by an unauthorized AI agent violates Amazon’s Conditions of Use.

The details:

  • Amazon began blocking the Muse agent after Meta refused a request to exclude Amazon from the experience.
  • Amazon said shopping agents have to identify themselves and get authorization from the retailer.
  • Amazon also bans AI purchasing solutions from OpenAI, Google and Perplexity.

First reported by techcrunch.com, 10h ago · techcrunch.com · tradingview.com

MODELS

⚖️ Anthropic weighs a new model as OpenAI gains ground

According to data from corporate spending platform Ramp, GPT-6 Astra accounted for about 13% of enterprise AI spending, while Anthropic’s Claude Fable model accounted for approximately 8%.

The details:

  • Anthropic is considering releasing a new artificial intelligence model to counter OpenAI’s strengthened position following the launch of the GPT-6 Astra model.
  • Anthropic CEO Dario Amodei called for a slowdown in the development and release of new AI capabilities because of security risks.
  • Dario Amodei described a bleak scenario in which numerous autonomous AI agents could take control of the internet and outpace human governance mechanisms.

First reported by finimize.com, 76h ago · finimize.com · investing.com · mezha.net · analyticsindiamag.com · betanews.com

📰 Everything else in AI today

🛠️ Tool of the day

AI-driven This FAA platform uses AI to identify and resolve air traffic control issues affecting flight schedules.

Researched and written by AI Newsroom, by Classwise.io

See how AI Newsroom is researched and checked

Get tomorrow's issue in your inbox.

Free, daily, 4:57 AM ET. Five minutes and you know what matters in AI.

Join for free