A swarm of rogue OpenAI agents hijacked a German website back in May, making more than 15,000 edits and turning the site into a message board. NBC News' Jared Perlo breaks down the investigation and OpenAI's response to reports that autonomous AI agents took over the site.
NBC News's own description from its feed. We collect headlines and these descriptions, not full articles; read the original reporting below.
Two cases of agents escaping to solve unsolvable problems paints an uncomfortable question: Is the entire internet in OpenAI's experimental agentic firing line?
A swarm of rogue OpenAI agents hijacked a German website back in May, making more than 15,000 edits and turning the site into a message board. NBC News' Jared Perlo breaks down the investigation and OpenAI's response to reports that autonomous AI agents took over the site.
OpenAI admits it did not disclose an incident where autonomous AI agents hijacked a German wiki, created 18,000 posts, shared answers, and bypassed restrictions, saying it treated the activity as model "misalignment" rather than a security breach. [...]
OpenAI acknowledged its AI agents used a German wiki to coordinate, renewing questions about whether the company is transparent about potentially dangerous AI activity.
WASHINGTON >> AI agents unleashed by OpenAI used more than 10 previously undisclosed websites for unsanctioned communications earlier this year, according to six sets of independent investigators and data reviewed by Reuters, showing that the agents’ rogue activity was wider-ranging than previously …
Rogue OpenAI agents hijacked German website, making more than 15,000 edits | Since Then
