Urgent.News

What's breaking now, across thousands of outlets.

Tuesday, 13 October 2026 Reading Room

AI

FBI arrests Saudi OpenAI employee for allegedly bugging San Francisco office

The FBI said on Wednesday it has arrested a Saudi national employed as an engineer at OpenAI on suspicions that he had bugged a conference room at the company's San Francisco office with a "sophisticated recording device".

france24.com ·

FBI arrests Saudi OpenAI employee for allegedly bugging San Francisco office

OpenAI terminated an employee after discovering a sophisticated recording device hidden in one of its conference rooms in San Francisco. The ChatGPT-maker announced the firing on October 14, stating that the individual planted the device a few days prior. The U.S. FBI is now investigating the matter, with the Justice Department confirming the arrest of the individual involved, Abdulrahman Alfozan, a 35-year-old Saudi Arabian national who had been working at OpenAI since July 2025.

OpenAI noted that there were no indications the employee improperly accessed or exfiltrated sensitive data. CCTV footage showed Alfozan near the conference room, and his productivity tool was used to track company meetings. Surveillance also revealed a chatbot account associated with his spouse, with prompts related to recording devices.

Another similar device was found in his bag, and an image of him appearing to look at the hidden device's location emerged. FBI Director Kash Patel emphasized the matter's significance for national security, while Director of National Intelligence Jay Clayton stated that a conviction for Alfozan could result in a five-year prison sentence, a $250,000 fine, and three years of supervised release.

Written by urgent.news from The Hindu - Sci-Tech's reporting — not their text. Machine-written — may contain errors; check the original before relying on it.

This story

This is one outlet's version. Read the fullest account.

Read the original at france24.com →

More in AI

Prompts don't enforce budgets. Signatures do.

A spending limit written in a prompt is not a control. It is a piece of text, sitting in the same context window as everything else the model reads, competing for the model's attention with the task…

  • Spending limits in prompts are not enforced by the model.
  • Real enforcement occurs through server-side checks.
  • Veyline performs scoped, signed authorization before transactions.

·dev.to

Log what your tools put in the context window before blaming the model

Ask an agent for one person's sales calls since February 1 and it can come back with calls that person wasn't on, even though the right ones were in its context.

  • Agent returned unrelated sales calls due to large context window
  • 30,000+ tokens consumed over a month and a half
  • Logging insertion point helps identify wrong claims

·dev.to

Test document extraction before an agent acts on it

An agent can follow its instructions perfectly and still pay the wrong amount, because the parser misread a table in the invoice. The agent never sees the PDF, only the extraction.

  • Agent relies on extracted data, not original PDF
  • Parser errors can cause workflow failure
  • Accuracy threshold advised at 95%

·dev.to

OpenAI and Anthropic call for new laws on artificial intelligence

OpenAI and Anthropic representatives told a UK parliamentary committee that voluntary pledges are insufficient and called for binding AI laws, statutory frameworks, and mandatory incident reporting.

  • OpenAI and Anthropic demand new AI laws in UK parliamentary hearing
  • Voluntary pledges insufficient, statutory frameworks and reporting required
  • Google DeepMind and Meta oppose new legal framework

·dev.to

More from Tuesday 13 October →