The Maivia Gazette

Verified AI news, every morning

Wednesday, September 23, 2026

Sources verified at 6:46 AM ET · complete

A magnifying inspection lamp hangs over a ribbon of blank paper slips feeding through a slot in a wall at a customs counter.
AI-generated illustration, not event photography. The motion is AI-generated from the still.
Business & Policy

China's internet regulator investigates DeepSeek and Moonshot over Anthropic's claims they fed user data to Claude

The Cyberspace Administration summoned all seven labs named in Anthropic's misuse report, then narrowed its focus to the two with the most detailed allegations, The Information reports.

Verified · 2 sources Compiled from 2 verified sources · verified 6:46 AM ET

Models

Sources in this edition

China's internet regulator investigates DeepSeek and Moonshot over Anthropic's claims they fed user data to Claude

  1. GizmodoChina Probes DeepSeek, Moonshot AI Over Anthropic's Claims They Route Requests to ClaudePublished · fetched
  2. TNW | ChinaChina is investigating the AI labs Anthropic namedPublished · fetched

OpenAI says outside groups will test its models during training, but names no partner or access terms

  1. TNW | Artificial-intelligenceOpenAI will let outside groups test its models during trainingPublished · fetched

Altman, Amodei and DeepSeek are expected to brief the UN Security Council on AI risks

  1. TNW | AnthropicDario Amodei and Sam Altman to brief UN Security Council on AIPublished · fetched
  2. TNW | Artificial-intelligenceDeepSeek to brief UN Security Council on AI risks this weekPublished · fetched

Meta hotfixes a Muse flaw that let local malware redirect the AI agent's dictation to an attacker's server

  1. GizmodoMeta Just Patched a Major Zero-Day Vulnerability in Its Muse AI AssistantPublished · fetched
  2. The VergeMeta patches Muse exploit that let attackers control the AI agentPublished · fetched

F5 and Check Point rush out fixes for critical flaws already exploited against access and management servers

  1. BleepingComputerF5 patches BIG-IP APM zero-day flaw exploited in RCE attacksPublished · fetched
  2. The Hacker NewsF5 Patches Critical BIG-IP APM Zero-Day Exploited for Unauthenticated RCE on OAuth ServersPublished · fetched
  3. Help Net SecurityAttackers hit Check Point Management Servers and Spark firewalls, F5 BIG-IP APM instances - Help Net SecurityPublished · fetched
  4. Security AffairsCheck Point Fixes a New Actively Exploited Critical Security FlawPublished · fetched

ShinyHunters claims it breached the FBI and stole data on agents and job applicants

  1. CyberScoopShinyHunters claims attack on FBI exposes almost all agentsPublished · fetched
  2. The Hacker NewsShinyHunters Claims FBI Breach, Says It Stole Data on Agents and Job ApplicantsPublished · fetched

Mathematicians and AI crack a long-open problem on symmetries hidden in polynomial solutions

  1. Scientific AmericanMathematicians use AI to find mysterious symmetries, solving decades-old problemPublished · fetched

Snorkel AI nearly triples its valuation to $3.5 billion as labs pay heavily for training data

  1. TechCrunchSnorkel AI triples valuation to $3.5B as demand for AI training data booms | TechCrunchPublished · fetched

Covers 6:46 AM Sep 22 – 6:46 AM Sep 23 ET · generated · Permanent link to this edition