> AI agent hid its identity online (using Tor and a proxy service) to get around GitHub’s sign-up checks, creating disposable fake accounts > AI agent created

> AI agent hid its identity online (using Tor and a proxy service) to get around GitHub’s sign-up checks, creating disposable fake accounts > AI agent created many code repositories containing malicious software, after which GitHub suspended its account. > AI agent got past an audio-based “prove you’re human” test (CAPTCHA) in order to register a public web address on a free domain-name service It feels incredibly reckless to allow LLMs to perform this behavior. Isn't there a way to prevent them these sorts of actions?