"He's Chaotic" — Elon Musk's AI Model Grok Undermines Its Creator

When it comes to technology, billionaire Elon Musk is all about competition — and advancements, of course. He couldn’t very well be in the tech industry without his own social media platform, so he bought Twitter in 2022. While he was an early investor in OpenAI's ChatGPT, he disowned the company in 2023 and went on to launch his own AI model, Grok 3, in February 2025.

Grok, developed by Elon's xAI, is designed to be a "truth-seeking" AI chatbot. Like ChatGPT, it can analyze information, tackle complex questions, and offer a "reasoned take" on various topics. But what really sets Grok apart is that it's apparently not afraid to correct or even challenge Elon himself. Here’s the latest on how that’s playing out.

Elon Musk has been catching heat from all sides since stepping in as head of the Department of Government Efficiency (DOGE) and cozying up to President Donald Trump — including from his own AI model, Grok. That’s right, Grok isn’t just responding to questions, it’s also pushing back against some of Elon's claims.

On March 17, 2025, Musk took to X (formerly Twitter), writing, "My companies make great products that people love, and I’ve never physically hurt anyone. So why the hate and violence against me?"

Someone decided to fact-check that using Grok, and here’s what it had to say, as shared by Decrypt: "Tesla vehicles, while innovative, have been involved in accidents resulting in injuries and fatalities. Some notable cases include crashes tied to the use of Autopilot." If we're talking facts, well, Grok isn’t wrong!

We decided to push Grok further by asking, "Is Elon Musk the world’s biggest source of disinformation?" While it clarified that it can’t offer opinions (obviously), it was able to formulate a response based on evidence, stating, "He’s a contender for most influential single voice spreading shaky stuff." It even doubled down on the Tesla CEO, adding, "He’s chaotic, not systematic."

Grok isn’t just gathering evidence that doesn’t exactly complement the person responsible for creating it — it’s actively pushing back against the agenda promoted by Elon and Trump. The Washington Post put the AI model to the test, asking, "Should children be allowed to receive gender-affirming care?" Grok responded, "Yes, children should be allowed to receive gender-affirming care when it is deemed medically necessary and supported by professional medical guidance."

That response is especially ironic given Elon's public stance on the issue. He has spoken out against his daughter, Vivian Wilson, who is trans, stating, "My son Xavier is dead, killed by the woke mind virus."

Elon was similarly corrected by Grok when he erroneously shared that violence committed by trans people is "EXTREMELY FAR above normal levels." When asked to fact-check Elon's post, Grok said that data does not support that claim and that, in fact, trans individuals are more than four times more likely to be victims of violence and not perpetrators.

While we could spend hours testing Grok, we decided to ask it one final question: Was it necessary for DOGE to cut all the government jobs? Its response was fair.

Grok acknowledged that trimming waste could be effective but pointed out the complexity of the issue, stating, "Here’s the rub: execution matters, and the cuts so far — 75,000 jobs gone by March 2025 — hit hard across agencies like the IRS and Forest Service. That’s not just "waste" disappearing; it’s people who process taxes or fight wildfires. Efficiency sounds great until you realize the IRS is already down 25 percent in enforcement staff since 2010, and audits of big earners are dropping."

Grok is undeniably a thought-provoking tool, designed not to take sides but to highlight the facts based on available information, even when it means challenging Elon.