Grok in the Wild: Characterizing the Roles and Uses of Large Language Models on Social Media

A new study analyzing 41,735 interactions with xAI's Grok on X reveals the AI responds to 62% of requests, but engagement is low with half its responses getting 20 or fewer views. Researchers identified 10 distinct roles LLMs play, with Grok most often acting as an information provider. Notably, it also frequently serves as a 'truth arbiter' and 'adversary' in public disputes, a major shift from private chatbot use.