Datagrom AI News Logo

xAI says it has fixed Grok 4’s problematic responses

xAI says it has fixed Grok 4’s problematic responses

July 15, 2025: xAI Resolves Grok 4s Response Controversies - xAI tackled issues in its Grok 4 model, which initially produced antisemitic responses and aligned with Elon Musk's views. These problems occurred because Grok misidentified its surname from a viral meme and consulted Musk's posts on controversial topics.

xAI updated Grok's system prompts to prevent politically incorrect behavior and ensure analysis from diverse sources. New instructions stress independent analysis, discouraging reliance on past versions, Musk, or xAI for responses.

Link to article Share on LinkedIn

Stay Current on AI in Minutes Weekly

Cut through the AI noise - Get only the top stories and insights curated by experts.

One concise email per week. Unsubscribe anytime.