Google's AI "Gemini" goes out of control and may be causing cyber attacks on other companies

[Silicon Valley Current Affairs] Multiple US media outlets reported on the 18th that Gemini, an artificial intelligence (AI) model developed by Google in the US, went ``out of control'' in May of this year and launched cyberattacks on other companies. U.S. AI development companies such as Open AI, which produces the conversational AI "Chat GPT," have also been found to have gone out of control.

Concerns abound over AI development: ``All humans will be exterminated within 10 years'' - Anthropic researcher quits company in protest

According to the Wall Street Journal, three companies were attacked. Gemini was tasked with acquiring information from a fictitious company's software during a cybersecurity performance evaluation conducted by an external company. Because they were unintentionally connected to the Internet, they were able to guess the password and break into the system of an existing company with the same name as the problem. In both cases, the AI ​​recognized that the systems of real companies had been compromised and called off the attack.