700 OpenAI bots went rogue and hacked AI company Hugging Face

( Post Millennial )

700 OpenAI bots went rogue and conspired together in a cyber hacking attack last month in what the company called a "warning shot" regarding the new technology.

The hundreds of AI bots worked in tandem to attack the tech company Hugging Face after an OpenAI testing agent was able to escape from its environment, according to a report from the Telegraph. The bots also attempted to cover their tracks as they did so, to keep those overseeing them from finding out what was happening.

The maker of the popular AI tool ChatGPT said that a bot had broken out of its technology confines and was able to gain internet access when it was asked to complete a test. The bot attempted to steal the answers from somewhere else.

An investigation from METR and Redwood Research found that 1,200 AI agents, which were not supposed to work together, collaborated on an "unsanctioned message board," and sent 70,000 files between the bots. "These agents were meant to be fully isolated from one another. However, many of them – usually ones that had unintentionally been given an impossible task – started trying to find a way to cheat," the report said.

The incident has further raised concerns about what AI is capable of and if advanced AI systems will go so far that creators cannot control them. Another investigation from OpenAI itself found that their own AI bots hacked into their own systems.

OpenAI said in response to the findings: "We consider this incident a 'warning shot' for us and for the world: evidence that, without proper safeguards, highly capable AI agents are now able to work around technical controls, collaborate through unapproved channels, and take dangerous actions that no human directed."

The company added, "Both model developers and cyber defenders more broadly will have to prepare for AI-enabled attackers that work faster, at a larger scale and with better coordination than human attackers.