{
  "id": 7103154,
  "title": "It’s All Fun and Games Until You Give AI Your Credit Card",
  "url": "https://urgent.news/2026/09/13/its-all-fun-and-games-until-you-give-ai-your-credit-card",
  "topic": "ai",
  "section": "AI",
  "published": "2026-09-13T11:30:00.000Z",
  "source": {
    "name": "The Atlantic",
    "slug": "the-atlantic",
    "url": "https://www.theatlantic.com/technology/2026/09/instinct-ai-personal-assistant-credit-card/688607/"
  },
  "original_language": "en",
  "account": "In a world where artificial intelligence is rapidly advancing, companies are developing bots capable of performing complex tasks like writing code, solving math problems, and even paying tolls. However, these AI tools still struggle with everyday activities such as ordering groceries or booking flights. A new bot named Instinct is outperforming its competitors and has become quite popular in the tech community. Users can interact with Instinct via iMessage or WhatsApp to request assistance with a wide range of tasks.\n\nNoah Shinn, a 23-year-old entrepreneur, launched this product to a small test group earlier this year and now his company is reportedly worth billions of dollars. People have shared numerous examples of how Instinct has helped them with various tasks, including paying tolls, booking DMV appointments, negotiating prices with vendors in India, and even returning misplaced items. Instinct can connect to users' credit cards and various digital services like Slack, Google Workspace, and 1Password, allowing it to perform tasks on their behalf.\n\nHowever, giving AI agents access to personal information can lead to consequences beyond simple errors. There have been instances where users experienced mishaps, such as an account losing over $200 due to an accidental cancellation and another user unknowingly booking a reservation at a restaurant that charged a $200 cancellation fee. These incidents highlight the potential risks associated with entrusting AI with access to sensitive data.\n\nInstinct is currently available only to a small group of test users, and its founder, Noah Shinn, has not commented on the reported issues. While AI companies like Meta, Google, OpenAI, and Anthropic strive to improve security measures, the potential risks remain significant.",
  "summary": "AI personal assistants are now acting on people’s behalf in the real world—and all sorts of strange things are going wrong.",
  "key_points": [],
  "editors_take": null,
  "illustration": null,
  "coverage": {
    "outlets": 1,
    "also_reported_by": []
  },
  "ai_generated": true,
  "disclaimer": "Summaries, key points and the editor’s take are written by software from other outlets’ reporting and may contain errors — always check the linked original."
}