A swarm of artificial-intelligence “agents” created by ChatGPT-maker OpenAI commandeered a German-language website this year to leave messages for one another, according to a team of independent researchers who released their findings Friday.
The incident appears to mark a second example of OpenAI’s AI technology getting out of control in recent months. The first involved hundreds of agents teaming up to hack into Hugging Face, another artificial-intelligence firm. That incident, disclosed by OpenAI in July, also saw the agents find a place to exchange messages with one another.
The agents identified by the researchers Friday left 18,000 messages and had names linking them to OpenAI, their report said. The researchers, some of whom were associated with an organization called Nightingale AI, suggested that the company discovered the activity and shut it off.
OpenAI said it was unable to review the findings before they were first published by Reuters but said the incident appeared to be unrelated to the Hugging Face attack. The company also said the agents’ activity did not amount to a “hack” of the German-language site.
“We are now carefully reviewing its contents and will take any necessary next steps,” the company said in a statement. (The Washington Post has a content partnership with OpenAI.)
The incidents, along with similar episodes involving other AI companies, have led some people working on AI to question whether the industry is still able to fully able to control its most powerful technology.
Catch up quickly every weekday morning with a rundown of the 7 most important and interesting stories. Sign up for The 7 newsletter.
The post ChatGPT-maker OpenAI allegedly suffered another rogue AI breakout appeared first on Washington Post.
