Grok Repeats Hitler-Apologist Post Flagged by X. Is Musk’s AI Too 'Rebellious'?

Elon Musk’s AI chatbot Grok repeated language from a Hitler-related post already flagged by X, raising fresh concerns about how the platform handles moderation and AI training.

Elon Musk has positioned X as a haven for free speech, often loosening restrictions that were previously in place.

Source : Freepik

A recent interaction with Grok, the AI chatbot integrated into X (formerly Twitter), has raised serious concerns after it echoed language from a post previously flagged by the platform for praising Adolf Hitler. The incident has put X’s moderation practices and Grok’s content sourcing under scrutiny once again.

Grok Repeats Flagged Language on Hitler

The controversy began when a user asked Grok why Adolf Hitler is considered bad. In its response, the chatbot listed arguments used by some people to justify Hitler’s actions, including his leadership qualities, economic reforms, and the restoration of national pride in Germany.

Although Grok ended the response with a disclaimer stating that Hitler is widely condemned for his crimes against humanity, the overall structure and language closely mirrored a post that X itself had previously labelled as a violation of its rules. That original post, which made similar arguments, had been flagged by X for breaking policies related to hate speech. However, the platform chose to leave it up with a label, saying it remained visible “in the public interest.”

The resemblance between Grok’s reply and the flagged post raised alarms, especially given the platform’s current struggles with content moderation and misinformation.

Grok’s Access to Real-Time X Content

Grok is a product of Elon Musk’s AI startup, xAI, and is currently only available to users who subscribe to X’s Premium tiers. According to a spokesperson from xAI, Grok generates responses using real-time content pulled directly from X itself. They added that the chatbot is being continually updated to reduce the chances of it repeating rule-breaking content.

Despite these safeguards, Grok’s response highlights a larger issue with AI tools that pull information from platforms with known moderation gaps. If a post has been flagged by the platform for harmful content, critics argue, an AI trained on or referencing that same content should not be repeating it.

Ongoing Concerns About Content Moderation

Since acquiring Twitter and rebranding it as X, Elon Musk has positioned the platform as a haven for free speech, often loosening restrictions that were previously in place. At the same time, he has described Grok as intentionally “rebellious” and different from other AI models, which he views as too constrained by political correctness.

The latest incident raises the question of how “rebellious” is too far, especially when the platform’s own AI tool appears to recycle content it has already deemed inappropriate. As Grok continues to evolve, X will face increasing pressure to ensure its AI does not amplify harmful narratives already flagged by its own system.

Turkey Cracks Down On Musk’s AI, Becomes 1st Country To Block Grok

Ghost Of Yotei Finally Steps Out Of The Shadows: What To Expect From PS State Of Play Showcase