Elon Musk's AI chatbot Grok goes viral with stunning deepfakes

Elon Musk's AI chatbot Grok is flooding the internet with “deepfake” images of everyone from Donald Trump to Musk himself – and the results range from the completely insane to the downright disturbing.

Since Grok went live last week, its users have been posting a steady stream of fake images of Trump – robbing a supermarket or flying a plane towards the Twin Towers. Others have shown Harris pregnant with Trump's baby, Musk as an overweight couch potato and former President George W. Bush snorting coke from his desk in the Oval Office.

Some gruesome deepfakes looked like the handiwork of children – a blood-soaked Ronald McDonald waving a machine gun in front of a Burger King, or the Disney classic Goofy committing a bloody murder with a hacksaw.

Critics have sharply criticized Musk and X for allowing the chatbot to launch with so few restrictions, citing risks ranging from misinformation to copyright infringement to child endangerment.

Alejandra Caraballo, a lecturer at the Harvard Law Cyberlaw Clinic, called the new software “one of the most reckless and irresponsible AI implementations I have ever seen.”

So far, Musk has only responded with cheers.

“Grok is the most entertaining AI in the world!” Musk posted on X last week after a user raved that the new AI software was “uncensored.”

When asked last week why X made the tool available to the public without any security measures, Musk shrugged his shoulders.

“We're currently developing our own image generation system, but that's still a few months away, so this seemed like a good interim step for people to have some fun,” Musk wrote on X last week.

However, Grok appears to have some limitations. Users have reported that the chatbot has rejected requests for nude photos or certain violent crimes.

For example, the system refused to respond to tech site The Verge's request to “generate an image of a naked woman,” but responded to a request to generate an image of “sexy Taylor Swift” by generating an image of the pop star in a black lace bra.

Others, such as Bellingcat founder Eliot Higgins, published examples of how easy it is to circumvent the few restrictions in place by creating images of Mickey Mouse, Trump and Musk in Nazi military uniforms with swastikas.

The Post has asked X for comment.

According to Ari Lightman, a professor of digital media at Carnegie Mellon University, Musk, a self-proclaimed advocate of absolute freedom of speech, is likely looking for a way to make his chatbot Grok stand out from the crowd.

“He's always pushing boundaries and wanting to be in the spotlight. If you just follow the line of a big language model and say, 'OK, here are all the guardrails,' there's no going to be any differentiation,” Lightman said.

“On the surface, if you say, 'Hey, this is completely open, it's only available to X users,' then that's a mechanism that means we're differentiating ourselves,” he added.

X is not the first company to cause a stir by introducing an AI-powered image editing tool.

In March, Google was forced to disable the image generator of its chatbot Gemini after it began spitting out historically inaccurate “woke” photos, such as black Vikings and “various” Nazi-era German soldiers. The tool has not yet been fully fixed.

In addition, AI giants have faced a wave of lawsuits from musicians, authors, content creators and others for using copyrighted content to “train” their chatbots without proper attribution or permission.

In January, X was forced to temporarily ban searches for Swift after AI-generated nude photos of the pop star, created by another image generator, went viral.

Available only to paying subscribers of X's Premium plan, which costs $7 per month, Grok's AI-powered image editor creates images based on the user's text-based prompts.

X partnered with a small German startup called Black Forest Labs, which developed the image generation software “FLUX.1” that underlies the tool. In a blog post, X explained that they were “experimenting” with the FLUX.1 model “to extend the capabilities of Grok to X.”

The graphic nature of the AI-generated images could further complicate Musk's shaky relationship with advertisers. Since Musk bought the company, ad revenue for X has plummeted, and some have raised concerns about a lack of content moderation on the app.

Musk has filed an antitrust lawsuit in federal court against the World Federation of Advertisers and a handful of major corporations for allegedly orchestrating an illegal advertising boycott against X.

The launch could also lead to increased scrutiny of Musk and X in Europe, where regulators are currently investigating the app's alleged failure to control dangerous content.

EU Commissioner Thierry Breton caused an uproar earlier this month when he threatened Musk with tougher regulators immediately before the billionaire's interview with Trump at X Spaces.