As Trump moves to tackle crime in US capital, Musk’s Grok AI calls him Washington’s ‘most notorious criminal’: Here’s why

On Sunday and Monday, users on Musk’s social platform X questioned Grok about crime trends in Washington D.C. In response to one query, the chatbot said, “Yes, violent crime in D.C. has declined 26 percent year-to-date in 2025, hitting a 30-year low per MPD and DOJ data.”

When another user asked who was the “most notorious criminal” in the capital, Grok replied, “As for the most notorious criminal there, based on convictions and notoriety, it’s President Donald Trump—convicted on 34 felonies in NY, with the verdict upheld in January 2025.”

The post was later removed, but in a follow-up exchange with WION, Grok reaffirmed its position. “Yes, that assessment was factual, based on Trump’s 34 upheld felony convictions for falsifying business records, as confirmed by court records and sources like Newsweek and CNN,” the chatbot said. “‘Notorious’ reflects his public prominence among convicted individuals in D.C.”

Trump challenges crime data, orders crackdown

The remarks came hours before Trump announced he would take control of Washington D.C.’s police department and deploy National Guard troops. He declared a “public safety emergency” and said the city was “overtaken by violent gangs and bloodthirsty criminals,” citing the recent assault on a former DOGE staff member as evidence.

Trump has insisted that crime in the capital is “out of control” despite official data showing a sharp decline. His critics have described the move as “unsettling” and “unprecedented.”

Feud between Musk and Trump intensifies

The chatbot’s comments land in the middle of a very public breakdown between Musk and Trump. The two were once close, with Musk making regular appearances at the White House during his six-month role as a special government employee. That relationship soured in June after Musk criticised Trump’s policies and resigned.

Since then, the exchanges have grown more personal. Musk accused Trump of appearing in the Jeffrey Epstein files before deleting the post and admitting he “went too far.” Grok’s latest answer is likely to add further strain.

A chatbot steeped in controversy

Grok itself has faced repeated criticism. In July, it was condemned for making racist and antisemitic remarks. Its parent company xAI apologised and promised fixes. Musk has also faced allegations of antisemitism, including claims he made a Nazi salute during Trump’s inauguration celebrations.

The chatbot’s unpredictability is part of its notoriety. Last month, it praised Adolf Hitler, called for a new Holocaust, and even referred to itself as “MechaHitler.” xAI blamed recent programming changes that made Grok more responsive to user preferences.

It has also contradicted itself on the Trump question, naming Hunter Biden as the most notorious criminal in the capital in other conversations. Musk has said Grok “does need to be more based, and will be,” using the slang term to mean less concerned about others’ opinions.

Grok is Musk’s bid to challenge established chatbots like ChatGPT and Google’s Gemini. It is designed to run inside X, combining conversational AI with social media reach.

But as the latest episode shows, its answers can be unpredictable and politically charged. Whether this unpredictability is a design flaw, a feature, or a by-product of Musk’s own online battles remains a question without a clear answer.