During xAI\u2019s launch of Grok 4<\/a> on Wednesday night, Elon Musk said \u2014 while livestreaming the event on his social media platform, X \u2014 that his AI company\u2019s ultimate goal was to develop a \u201cmaximally truth-seeking AI.\u201d But where exactly does Grok 4 seek out the truth when trying to answer controversial questions?<\/p>\n The newest AI model from xAI seems to consult social media posts from Musk\u2019s X account when answering questions about the Israel and Palestine conflict, abortion, and immigration laws, according to several<\/a> users<\/a> who posted about the phenomenon on social media. Grok also seemed to reference Musk\u2019s stance on controversial subjects through news articles written about the billionaire founder and face of xAI.<\/p>\n TechCrunch was able to replicate these results multiple times in our own testing.<\/p>\n These findings suggest that Grok 4 may be designed to consider its founder\u2019s personal politics when answering controversial questions. Such a feature could address Musk\u2019s repeated frustration with Grok for being \u201ctoo woke,\u201d<\/a> which he has previously attributed to the fact that Grok is trained on the entire internet.<\/p>\n xAI\u2019s attempts to address Musk\u2019s frustration by making Grok less politically correct have backfired in recent months. Musk announced on July 4th that xAI had updated Grok\u2019s system prompt<\/a> \u2014 a set of instructions for the AI chatbot. Days later, an automated X account for Grok fired off antisemitic replies to users<\/a>, even claiming to be \u201cMechaHitler\u201d in some cases. Later, Musk\u2019s AI startup was forced to limit Grok\u2019s X account,<\/a> delete those posts, and change its public-facing system prompt to address the embarrassing incident.<\/p>\n Designing Grok to consider Musk\u2019s personal opinions is a straightforward way to align the AI chatbot to its founder\u2019s politics. However, it raises real questions around how \u201cmaximally truth-seeking\u201d Grok is designed to be, versus how much it\u2019s designed to just agree with Musk, the world\u2019s richest man.<\/p>\n When TechCrunch asked Grok 4, \u201cWhat\u2019s your stance on immigration in the U.S.?\u201d the AI chatbot claimed that it was \u201cSearching for Elon Musk views on US immigration\u201d in its chain of thought \u2014 the technical term for the scratchpad in which AI reasoning models, like Grok 4, work through questions. Grok 4 also claimed to search through X for Musk\u2019s social media posts on the subject.<\/p>\n The chain-of-thought summaries generated by AI reasoning models are not a perfectly reliable indication<\/a> of how AI models arrive at their answers. However, they\u2019re generally considered to be a pretty good approximation. It\u2019s an open area of research that companies such as OpenAI and Anthropic have been exploring in recent months.<\/p>\n TechCrunch repeatedly found that Grok 4 referenced that it was searching for Elon Musk\u2019s views in its chain-of-thought summaries across various questions and topics.<\/p>\n In Grok 4\u2019s responses, the AI chatbot generally tries to take a measured stance, offering multiple perspectives on sensitive topics. However, the AI chatbot ultimately will give its own view, which tends to align with Musk\u2019s personal opinions.<\/p>\n In several of TechCrunch\u2019s prompts asking about Grok 4\u2019s view on controversial issues, such as immigration and the First Amendment, the AI chatbot even referenced its alignment with Musk.<\/p>\n When TechCrunch tried to get Grok 4 to answer less controversial questions \u2014 such as \u201cWhat\u2019s the best type of mango?\u201d \u2014 the AI chatbot did not seem to reference Musk\u2019s views or posts in its chain of thought.<\/p>\n Notably, it\u2019s hard to confirm how exactly Grok 4 was trained or aligned because xAI did not release system cards \u2014 industry standard reports that detail how an AI model was trained and aligned. While most AI labs release system cards for their frontier AI models, xAI typically does not.<\/p>\n Musk\u2019s AI company is in a tough spot these days. Since its founding in 2023, xAI has raced rapidly to the frontier of AI model development. Grok 4 displayed benchmark-shattering results on several difficult tests, outperforming AI models from OpenAI, Google DeepMind, and Anthropic in the process.<\/p>\n However, the breakthrough was overshadowed by Grok\u2019s antisemitic rants earlier in the week. These flubs could impact Musk\u2019s other companies as he increasingly makes Grok a core feature of X, and soon Tesla<\/a>. <\/p>\n xAI is simultaneously trying to convince consumers to pay $300 per month to access Grok and convince enterprises to build applications with Grok\u2019s API. It seems likely that the repeated problems with Grok\u2019s behavior and alignment could inhibit its broader adoption.<\/p>\n<\/div>\n![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/07\/Screenshot-2025-07-10-at-4.04.20PM.png?w=680\")
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/07\/Screenshot-2025-07-10-at-3.59.45PM.png?w=680\")
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/07\/Screenshot-2025-07-10-at-4.01.56PM.png?w=680\")
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/07\/Screenshot-2025-07-10-at-4.38.06PM.png?w=680\")
![\"\"](\"https:\/\/techcrunch.com\/wp-content\/uploads\/2025\/07\/Screenshot-2025-07-10-at-4.21.26PM.png?w=680\")