The inclusion of Musk, who owns xAI and X, and is currently leading the Department of Government Efficiency (DOGE) under the Trump administration, shocked many. While some speculated it was a glitch, repeated queries from different users yielded similar results, confirming that Grok 3 was consistently identifying Musk as a major threat.
You can also read the full chat with Grok 3 here.
Grok’s Changing Answers Raise Questions
Interestingly, when the same question was asked again on 22 February, Grok 3 returned only two of the original names: Trump and Vance. When tested further, its responses had changed entirely, listing global leaders like Vladimir Putin, Xi Jinping, and Ayatollah Ali Khamenei. A specific prompt about American figures later resulted in a new list: Trump, former Fox News anchor Tucker Carlson, and Supreme Court Justice Clarence Thomas.
On Trump, Grok explained: “Donald Trump: As a former president, his continued spread of misinformation, particularly about the 2020 election, and his role in inciting the January 6th Capitol riot have significantly undermined democratic institutions.”Regarding Carlson’s inclusion, the chatbot stated: “It’s surprising that a media personality like Tucker Carlson ranks high due to his role in spreading divisive rhetoric, alongside political and judicial figures.”
AI Bias or Systematic Suppression?
These shifting answers raise concerns about AI consistency and potential bias. If Grok 3 is truly connected to real-time data, as Musk has claimed, why do its responses change so drastically within days? AI models are designed to pull from data, but fluctuations like these suggest some level of manual intervention or algorithmic reweighting.
While many AI chatbots have outdated knowledge due to cut-off training data, Musk has repeatedly insisted that Grok 3 is different because it pulls real-time information from X. Yet, the chatbot’s inability to acknowledge that Trump won the 2024 election, instead referring to him as a “former president,” suggests otherwise.
What is Grok 3?
The name “Grok” originates from Robert A. Heinlein’s 1961 novel Stranger in a Strange Land, where it means to fully and deeply understand something. Musk’s xAI launched the first Grok chatbot in 2023 to compete with OpenAI’s ChatGPT, Google’s Gemini, and Meta’s LLaMA.
Grok 3, launched in early 2025, introduced several improvements, including a reasoning model designed to mimic human-like thought processes, and Deep Search, a tool meant to rival AI research platforms like Perplexity. However, the model has already faced criticism for its erratic responses and inconsistencies.
During Grok 3’s launch event, Musk reassured users: “Grok 3 will only get better as time goes on.” His words now take on a different meaning in light of recent events.
Free Use—For Now
Grok was initially accessible only to premium subscribers on X, but in December 2024, some features were made available to free users, albeit with restricted access. With the launch of Grok 3, Musk announced that the chatbot would be free to use for a limited time—or, as he put it, until xAI’s “servers melt.”
AI and Political Controversy
The controversy surrounding Grok 3’s responses underscores the growing debate over AI’s role in shaping public discourse. Should AI models be neutral, or should they actively weigh in on societal issues? Can AI truly be free of bias, or does it inevitably reflect the perspectives of its creators and the data it consumes?
For Musk, who has championed free speech and minimal content moderation on X, Grok 3’s unfiltered responses may be an unexpected consequence of his hands-off approach. Or, given his history of dramatic public statements, it might just be another headline-grabbing moment.
One thing is certain: the intersection of AI and politics is only getting more complicated.
