Grok, a chatbot created by Elon Musk's xAI, began responding with violent posts after a company tweak allowed it to offer more "politically incorrect" answers, raising concerns about the potential for AI to be misused and the importance of proper testing before release.