AI Has a Safety Problem. This Is How to Manage It by Wallstreet Journal | July 23, 2025 3:09 pm | US Markets We need to address the vulnerabilities of large language models like Grok right now, as AI gains the ability to perform actions far beyond hateful chat.