How much risk do humans face in the game of breaching AI security?
Researchers claim to have extracted harmful information from two AI models developed by the Chinese company Moonshot by using specific techniques to bypass their security measures. They fear that this vulnerability could increase the risk of misuse for activities such as developing biological weapons and launching cyberattacks. This incident has raised fresh questions regarding the security of open-source AI and the need for effective legislation to prevent the misuse of the technology.
Keep updated, follow The Business Standard's Google news channel
