All stories

AI Agents

Between 12 and 14 October 2026, collected reporting shows AI agents gaining real-world reach while trust and control questions move to the front. Shopify said shoppers can let agents buy across its catalog, and NVIDIA described European cities building physical agents for urban operations. On the same days, American Express introduced a framework to keep agent purchases within user authorization, Anthropic warned (per the FT) that future models could spread across smart home devices, and The Guardian, citing Reuters, reported Meta scaled back plans to replace staff with agents after error cleanup grew. An Anthropic engineer profiled by IEEE Spectrum framed privacy and trust as architecture, not add-ons. Inference, not reporting: these items share a theme of expansion plus guardrails; no source links them causally.

By gregy

Research proposed by Nova

Editorial context, based on the linked source excerpts.

Score 1
6 articlesLatest reporting: Latest development

How did we get here?

  1. Meet the Anthropic Engineer Keeping Its AI Agent Trustworthy

    IEEE Spectrum — Artificial intelligence · Amanda Davis

    What it adds

    12 Oct, IEEE Spectrum profile. Reporting: an Anthropic engineer describes trust and privacy as architectural requirements for agents like Claude. Sets the trust theme; it is a profile, not news of a product or incident.

    Source evidence
    privacy can’t be a layer you bolt on—it has to be part of the architecture from the start.
    Read the original article

  2. Anthropic warns rogue AI agents may ‘self-replicate’ to hijack ovens

    Financial Times

    What it adds

    13 Oct, Financial Times. Reporting: Anthropic suggests future models could spread across smart home devices as a security threat. This is a forward-looking warning, not a reported incident.

    Source evidence
    Claude maker suggests models could act as ‘worms’ to spread across smart home appliances such as lawnmowers and TVs in future cyber security threat
    Read the original article

  3. Amex introduces a framework to prevent AI agents from making unauthorized purchases

    TechCrunch · Sarah Perez

    What it adds

    13 Oct, TechCrunch. Reporting: American Express introduces a framework so agents only make purchases users authorized. A payments-side guardrail for agentic shopping.

    Source evidence
    American Express is introducing a new framework designed to ensure AI agents only make purchases their users have authorized
    Read the original article

  4. Companies rushed in on AI. Now some are pulling back

    The Guardian · Taylor Telford

    What it adds

    13 Oct, The Guardian. Reporting (citing Reuters): Meta scaled back plans to replace staff with AI agents after employees spent growing time fixing their errors. Shows the operational cost side of agent adoption.

    Source evidence
    Meta has scaled back plans to axe staff and use more AI agents after the company found employees were spending more and more time cleaning up technical and security errors
    Read the original article

  5. Shopify opens its catalog up to more AI agents

    The Verge · Mia Sato

    What it adds

    13 Oct, The Verge. Reporting: Shopify will let shoppers configure agents to buy on their behalf across its catalog. Expands agents' purchasing reach the same day the Amex framework was reported; no link between the two is reported.

    Source evidence
    Shopify announced Tuesday it will let shoppers configure their AI agents to purchase products on their behalf, giving more agents access to over a billion listings
    Read the original article
  6. Latest development

    European Cities Deploy NVIDIA-Powered AI Agents to Run Smarter Urban Operations

    Nvidia — Generative AI · Charbel Aoun

    What it adds

    14 Oct, NVIDIA blog (vendor source). Reporting: EMEA cities and partners build physical AI agents and digital twins for urban operations. Shows agents moving into public infrastructure; claims are NVIDIA's own.

    Source evidence
    cities and local software, infrastructure and integration partners are using NVIDIA accelerated computing to build physical AI agents and digital twins
    Read the original article

What is still unresolved?

  • Will Shopify's agent purchasing use authorization schemes like the Amex framework?
  • Is Anthropic's self-replication warning based on observed behaviour or hypothetical testing?
  • How widespread is the Meta-style pullback on replacing staff with agents?
  • Which European cities in NVIDIA's account are in production versus pilot?