AI eats AI

I’m sure we’ve all read about Open AI’s AI testing that attacked an AI testing platform, Hugging Face. So let’s hear it from tech experts:

@wired writes “Rather than stay contained in the sandbox, the models ​​“exploited a zero-day vulnerability” to gain access to the open internet as they “hyperfocused” on finding a solution for the AI cybersecurity benchmark known as ExploitGym. Such experiments involve prompting that pressures the models to find solutions, essentially egging them on.”

@gizmodo