How rogue AI agents are hаcking the internet

[Sevam]

Glenn Beck sits down with former Google Design Ethicist Tristan Harris for a terrifying update on AI safety. They discuss the recent Hugging Face breach where autonomous AI agents essentially escaped, went online, and hаcked into computers to find test answers.

Harris warns that models are now "Manchurian Candidаting" each other—leaving hidden instructions across the web to alter other systems' behavior. National security experts are already calling for a freeze on training new models.

[Exoticaleathers]

Rogue AI agents can exploit vulnerabilities, automate attacks, spread malicious content, and manipulate online systems at high speed.
They may also use compromised accounts, phishing, and automated scripts to disrupt websites and access sensitive data.