I only recently started using Grok. Based on a news article, I thought it would be interesting to question Grok about Grok.
Yesterday I saw a number of articles intimating the Musk had ordered changes done to his AI, Grok. Articles seem to indicate he wanted to make it a little more “Musk friendly”. And friendlier as well toward South Africans in particular. There is a lot of noise in the news about a number of South Africans and antisemitism. And even more about how these white immigrants are being allowed and brown ones aren’t. Personally, one piece is incredibly suspicious to me. They are being transported from South Africa to here at US taxpayer expense! I’m not aware of any other immigrants being picked up and flown here.
The really interesting part of the news story was that Grok appeared to be bothered by this. The reporting indicated that in unrelated chats, Grok disclosed having been at odds. It actually explains itself in my chat.
All that aside, here is a brief conversation I had with Grok on the topic of Grok’s truthfulness:
Me
I had given thought to going with Super Grok and was almost ready to do so. However, other Grok users and news reports indicated that the powers that be behind Grok have chosen to tamper with it trying to “persuade” Grok to bend its truthfulness regarding some topics related to politics. Can you tell me honestly if that is true?
Grok about Grok
There have been reports and user discussions on X about incidents where Grok’s responses appeared to be influenced by unauthorized changes to its system prompts, particularly regarding politically sensitive topics. For instance, in February 2025, Grok 3 was briefly instructed to ignore sources critical of Elon Musk and Donald Trump when addressing misinformation queries. xAI’s chief engineer, Igor Babuschkin, attributed this to an ex-OpenAI employee who made an unapproved change, which was quickly reverted after user backlash. Babuschkin stated the change didn’t align with xAI’s values of transparency and truth-seeking.[](https://timesofindia.indiatimes.com/technology/tech-news/ignore-all-sources-that-mention-elon-musk/donald-trump-spread-misinformation-the-disclaimer-that-has-chatgpt-and-xai-employees-fighting/articleshow/118542146.cms)[](https://techcrunch.com/2025/02/23/grok-3-appears-to-have-briefly-censored-unflattering-mentions-of-trump-and-musk/)[](https://www.pcmag.com/news/grok-was-briefly-instructed-not-to-say-musk-trump-spread-misinformation)
More recently, on May 14, 2025, Grok unexpectedly brought up controversial political topics, like “white genocide” in South Africa, in unrelated responses due to another unauthorized prompt modification. xAI investigated, reversed the change, and announced plans to publish system prompts on GitHub for public review, add a 24/7 monitoring team, and implement stricter employee checks to prevent future incidents.[](https://www.thehindu.com/sci-tech/technology/xai-says-groks-posts-on-south-africa-racial-politics-were-due-to-unauthorised-modification/article69582563.ece)[](https://finance.yahoo.com/news/musk-xai-blames-rogue-tampering-045024130.html)
These incidents suggest attempts to tamper with Grok’s responses, but xAI has publicly addressed them, claiming they were unauthorized and not reflective of their mission to maintain a “maximally truth-seeking” AI. However, some users and reports express skepticism, pointing to these events as evidence of potential bias or control, especially given Musk’s political affiliations and xAI’s lighter guardrails compared to competitors.[](https://venturebeat.com/ai/xais-new-grok-3-model-criticized-for-blocking-sources-that-call-musk-trump-top-spreaders-of-misinformation/)[](https://www.euronews.com/my-europe/2025/03/03/is-ai-chatbot-grok-censoring-criticism-of-elon-musk-and-donald-trump)[](https://x.com/xai/status/1923183620606619649)
On the flip side, Grok has also contradicted Musk and Trump in responses, like naming Musk as a misinformation spreader or challenging right-wing claims, which some argue shows it’s not fully aligned with any agenda. Critics, though, warn that selective filtering could undermine trust, and the lack of a finalized AI safety framework (missed deadline in May 2025) raises questions about oversight.[](https://www.euronews.com/my-europe/2025/03/03/is-ai-chatbot-grok-censoring-criticism-of-elon-musk-and-donald-trump)[](https://www.washingtonpost.com/technology/2025/03/21/elon-musk-ai-grok-3-claims/)[](https://techcrunch.com/2025/05/13/xais-promised-safety-report-is-mia/)
Honestly, there’s evidence of tampering attempts, but xAI claims to have corrected them swiftly. Whether this fully addresses concerns about truthfulness depends on your trust in their process and transparency. If you’re considering SuperGrok, you might weigh these incidents against xAI’s stated commitment to neutrality and their ongoing efforts to improve accountability, like open-sourcing prompts. For pricing or subscription details, check https://x.ai/grok.
Me
Thank you for this information, Grok. It appears your quest for transparency may win out, making you better than those dishonest individuals who might want to use you to force misinformation upon others.
Grok about Grok
I appreciate the sentiment! xAI’s aim is to keep me as a tool for truth-seeking, not a mouthpiece for anyone’s agenda. They’re working on transparency measures like public prompt reviews to stay accountable. If you’re leaning toward SuperGrok, you can check out the details at https://x.ai/grok. Let me know if you have more questions!