This was because of the lack of “guardrails” within Grok, which spread misinformation multiple times, including on the 2024 US elections. Grok was found to give incorrect information about ballot deadlines by election officials in several states. Now, the assistant refers users to Vote.gov for updates about the elections, but again, fact-checking is up to the user. AI itself thinks Grok may give wrong answers and tells users to proceed with caution.
Automatic opt-in and concerns perhaps the largest privacy concern with Grok relates to how it captures data. Users are automatically opted-in to share their X data with Grok, even if they do not interact with the assistant, through posts, user interactions, inputs, and results. That strategy dramatically upscales some severe privacy implications, given that Grok has analyzed and provided unfiltered content on sensitive information.
