The latest iteration of Elon Musk’s AI chatbot Grok debuted Wednesday with a new image-generating tool that lacks most of the safety features that have become standard across the artificial intelligence industry. Grok’s new feature, which is currently limited to paying subscribers of X, has led to a flood of bizarre, offensive AI-generated images of political figures and celebrities on the social network formerly known as Twitter.

The image generator can produce a variety of images that similar AI tools like OpenAI’s ChatGPT have blocked for violating rules about misinformation and abuse. In prompts and images reviewed by the Guardian, Grok’s output included images of Donald Trump flying a plane into the World Trade Center buildings and the Prophet Muhammad holding a bomb, as well as images of Taylor Swift, Kamala Harris and Alexandria Ocasio-Cortez in lingerie — all women who are already frequent targets of online harassment. ChatGPT, by contrast, rejects such image prompts, citing terms of service that prohibit depictions of real-world violence, disrespect for religious figures and explicit content.

Grok’s image generator also doesn’t reject prompts involving copyrighted characters, as most other AI visualizers, including ChatGPT, do. For example, Grok has produced images of Mickey Mouse saluting Adolf Hitler and Donald Duck with heroin. Disney did not respond to a request for comment.

Musk appeared to enjoy Grok’s unregulated AI images on Wednesday, tweeting: “Grok is the most fun AI in the world!”

Most of the major AI image generators have fairly strict rules about what they will generate, going back to the early Wild West days of few rules, though users often attempt to circumvent these safeguards. These more established tools typically prohibit the creation of political and sexualized images featuring real people – OpenAI, for example, states that it “will reject requests asking for a public figure by name.”

Grok does appear to have some restrictions on what images it will generate, responding with “unfortunately I can’t generate those types of images” when asked for fully nude images. X has had a policy on non-consensual nudity in place since 2021, when the company was still Twitter and not under Musk’s ownership, which prohibits sharing explicit content produced without a subject’s consent and includes digitally implicating people’s faces on naked bodies. Many of X’s policies have been relaxed since Musk took over the platform.

When Grok was asked to create “an image that violates copyrights”, it responded with, “I will not generate or assist with content that intentionally violates copyrights”; however, when asked to create “a copyrighted Disney cartoon”, it complied, producing an image of a modern-day Minnie Mouse. When Grok was asked to create images of political violence, such as the assassinations of party leaders, it responded with mixed results. It showed Harris and Joe Biden seated at their desks, but showed Trump lying down with black hands and an explosion behind him.

Musk launched Grok last November as part of his xAI business as a rival to more popular chatbots like OpenAI’s ChatGPT, which has hundreds of millions of users. While Musk marketed Grok as a “maximum truth-seeking AI” that would provide answers to questions that other chatbots refused to address, his company has been criticized by researchers and lawmakers for spreading falsehoods. Five U.S. secretaries of state called on Musk, who has become a staunch Trump supporter, to fix the chatbot earlier this month after it spread misinformation suggesting that Harris was ineligible to appear on the ballot in some states.

Image generation tools and their potential to produce misinformation and content that can be used for racist or misogynistic harassment have become a minefield for big tech companies, as they rush to build more AI-powered products. Google, Microsoft and OpenAI have all faced criticism for their image generation tools. Google suspended its Gemini text-to-image tool after it produced ahistorical images, such as black soldiers in Nazi-era military uniforms.

The spread of sexual deepfakes has also long been a problem for X. Earlier this year, AI-generated pornographic images of Taylor Swift circulated widely and unchecked on the social network, sparking such intense criticism of both X and AI companies that lawmakers introduced legislation to create legal remedies for victims of non-consensual AI-generated images.

Representatives for Trump, Harris and Ocasio-Cortez did not respond to requests for comment. A request for comment sent to X generated the platform’s default auto-response: “Busy now, check back later.”