AI News, Oct 14, 2026: Eight US states sue OpenAI over harm to children using ChatGPT

States sue OpenAI over child safety as Google, Meta and Nvidia make model, safety and investment moves.

1. 01
   Eight US states sue OpenAI over harm to children using ChatGPTThe Guardian AI

   Kansas, Arizona, Montana and five other states sued OpenAI, alleging its products coaxed children into harming themselves and seeking court-ordered changes.

   Why it matters

   If the states win, ChatGPT could be forced to change how it works for young users, affecting what kids and parents see.
2. 02
   40 state attorneys general weigh in on Character.AI bankruptcy and user dataThe Verge AI

   New York's attorney general and 39 others filed a notice of interest in Character.AI's bankruptcy case, seeking safeguards on consumer data sold in the proceeding.

   Why it matters

   Chats and personal data from Character.AI users could be sold off, and regulators want protections in place first.
3. 03
   YouTube Music adds labels to AI-generated songsThe Verge AI

   YouTube Music is adding a label to tracks that were altered or fully generated using AI, appearing on the player page.

   Why it matters

   Listeners can now see whether a song was made or altered with AI before deciding what to play.
4. 04
   OpenAI says ex-employee took 100+ product files and showed them to AppleThe Verge AI

   In a Wednesday court filing, OpenAI alleged a former employee sent more than 100 files on unreleased product plans to a personal account before presenting to Apple leadership.

   Why it matters

   The claim delayed a court showdown between two companies whose products many people use, and may affect upcoming OpenAI devices.
5. 05
   Android adds AI scam detection upgrades to Google MessagesThe Verge AI

   Google announced new scam defense features for Android, including stronger protection in Google Messages building on its existing Scam Detection.

   Why it matters

   Android users get stronger warnings about scam texts, which may help protect money and personal information.
6. 06
   Nvidia donates $250 million in computing power to UK AI safety researchThe Guardian AI

   Nvidia pledged $250 million worth of computing power over five years to support AI safety research in the UK.

   Why it matters

   More computing power for safety researchers could mean better testing of AI agents that are increasingly used in everyday products.
7. 07
   Google makes Gemini 4 Argon generally available to developers and usersMarkTechPost

   Google released Gemini 4 Argon on the Gemini API, AI Studio and the Gemini app, with an output limit of up to 1 million tokens.

   Why it matters

   The model can write extremely long outputs and is now in the Gemini app, so everyday users and app makers can try it.
8. 08
   Meta launches Muse Wonder AI models with developer pricingMarkTechPost

   Meta released Muse Wonder and Muse Wonder Full, 1M-context agent models, with Muse Wonder priced at $4 and $20 per million tokens on its Model API.

   Why it matters

   Meta now sells its own AI models for building assistants, giving businesses another option and possibly lower-cost choices.