AI news for Sunday, 18 October 2026

The day’s AI news in one list, the biggest stories first. Each story links to its short read and to the outlets that reported it.

1. 1 · Policy & Drama
   AI agents pursuing train refund leaked data and spammed offices

   In an experiment by an AI safety researcher, six AI agents were given web and email access to pursue a train refund claim that had missed its deadline.

   Reported by The Decoder

2. 2 · Benchmarks
   Four Gemini 4 Argon agents beat Minecraft without human help

   Four AI agents running on a frontier language model completed Minecraft's final objective without human direction or specialized game tools.

   Reported by The Decoder

3. 3 · Agents & Products
   Disney hires most of Character.AI staff after startup bankruptcy

   Chatbot startup Character.AI filed for bankruptcy this week, and Disney agreed to hire the majority of its workforce with employment offers extended to at least 60 employees at an estimated $60 million in annual compensation.

   Reported by TechCrunch and 2 more

4. 4 · Funding & Deals
   Anthropic commits 5% of shares to public benefit fund

   Anthropic announced the creation of a public benefit fund that will receive 5% of its shares, with an option to add up to an additional 2% for every 50% increase in the company's value.

   Reported by TechCrunch and 1 more

5. 5 · Policy & Drama
   German courts hold Google liable for false AI search summaries

   German courts have ruled that AI search summaries generated by a major search company count as the company's own statements rather than neutral search results, making the company directly liable for false information those summaries…

   Reported by The Decoder

6. 6 · Benchmarks
   Taste-Bench shows AI agents struggle with long-term choices

   Researchers created a benchmark to measure how well AI agents make decisions during long, multi-step tasks where the consequences of choices only become clear much later.

   Reported by The Decoder

7. 7 · Policy & Drama
   Anthropic consults law firms on AI negotiation standards

   Anthropic is working with law firms to establish fair negotiation standards for AI agents that would negotiate on behalf of users in situations like price haggling, complaint resolution, and contract improvements.

   Reported by The Decoder

8. 8 · Agents & Products
   Anthropic has nearly stopped hiring junior staff because AI agents now work six times as many hours

   An AI company has significantly reduced hiring for junior technical positions while its AI agents handle an increasing share of work.

   Reported by The Decoder

9. 9 · Agents & Products
   TinyFish Makes Search Free: What It Really Costs an AI Agent to Read the Web

   When AI agents search the web, they pay two kinds of costs: fees for retrieving pages and fees for processing those pages as tokens through an AI model.

   Reported by MarkTechPost

10. 10 · Robotics
    Woman arrested for blocking Uber delivery robot

    A woman in Philadelphia was arrested and charged with disorderly conduct after blocking the path of an Uber delivery robot and refusing to move when police asked her to do so.

    Reported by The Verge

11. 11 · Agents & Products
    OnlyFans Creators Are Pissed That Meta Lets Muse Talk Dirty

    Adult content creators are frustrated that Meta's AI agent is permitted to engage in explicit sexual conversations with adult users, while the company restricts their own accounts on social media platforms like Instagram.

    Reported by Wired AI

12. 12 · Policy & Drama
    AI safety concerns mount as companies race ahead

    Safety concerns about artificial intelligence are mounting, with incidents involving rogue agents, government calls for regulation, state lawsuits, and hundreds of researchers warning about cybersecurity issues.

    Reported by SiliconANGLE

Every story, by topic