European leaders are now fretting about Vladimir Putin in much the same way that everyone worries these days about artificial intelligence. Until recently, AI agents remained within the virtual spaces where they’d been developed. Rather than free range, they were confined to their “sandboxes.” But there have been a number of cases of these AI agents lying, cheating, and hacking their way out of their designated spaces. Now, even AI experts are talking about the not-insignificant odds that the technology they designed will run amok and kill all of humanity. More









