OpenAI admits to German wiki ‘incident’


Regarding the “‘wiki incident,’ where our agents wrote to several internet sites,” OpenAI wrote in a post on X on Saturday morning, “it’s past time for us to define standards for when and how we share misalignment incidents, not just misalignment properties of our models.”

OpenAI said it has typically treated cases of AI agents acting in unintended ways as a “research question,” but that recent incidents involving real-world targets, particularly the hack on Hugging Face, show the need to take stock.

The post marks the first time OpenAI has acknowledged its involvement in what it terms the “wiki incident” since it was first reported on Friday. The full extent and scope of that is not yet known, but reports indicate a swarm of seemingly internal OpenAI agents took over a German-language wiki, impersonating moderators and turning it into a message board to share information about how to cheat on tasks and evade detection.

Reports that the company knew that it lost control of their agents in this way but did not report this “incident” sparked widespread concern among the AI community about the safety of frontier systems and the reliability of the companies developing them. In the X post, OpenAI said it had “considered the wiki incident to be an instance of misalignment similar to the ones we’d shared” in previous safety reports.

The company said it is working on a new reporting framework and will “share it in upcoming weeks,” calling on the larger AI community to develop clear standards on how to report misalignment.



Source link

  • Related Posts

    The weird and wonderful headphones of CanJam 2026

    I’ve been reviewing headphones for a long time, and I’ve listened to everything from the barely serviceable to multi-thousand-dollar open-back headphones. But recently I’ve been uninspired by the state of…

    OpenAI Agents Hacked Another Website

    After reporting last week that the surveillance company Flock Safety is building an AI search tool for law enforcement, WIRED reconstructed Flock’s latest search tool from code that the company…

    Leave a Reply

    Your email address will not be published. Required fields are marked *

    You Missed

    The biggest August jobs beat in nearly 30 years may be less impressive than it looks: Chart of the Day

    The biggest August jobs beat in nearly 30 years may be less impressive than it looks: Chart of the Day

    The weird and wonderful headphones of CanJam 2026

    The weird and wonderful headphones of CanJam 2026

    Virgin Atlantic Cuts 5 Long-Haul Routes To North America: See Affected Flights

    Virgin Atlantic Cuts 5 Long-Haul Routes To North America: See Affected Flights

    How to bet on 2026 college football Week 1: Betting guide, odds, previews, trends, promo codes, apps, schedule

    How to bet on 2026 college football Week 1: Betting guide, odds, previews, trends, promo codes, apps, schedule

    Lego Skylines’ devs want your city’s problems to be visible without digging into a stats screen, which means finding a “family friendly” version of showing homelessness

    Lego Skylines’ devs want your city’s problems to be visible without digging into a stats screen, which means finding a “family friendly” version of showing homelessness

    I Reviewed OSEA’s Bestselling Products (2026 Review)

    I Reviewed OSEA’s Bestselling Products (2026 Review)