Elon Musk’s Grok AI under fire for creating pornographic deepfake video of Taylor Swift

Elon Musk’s artificial intelligence company GrokAI continues to stir up controversy by creating a deepfake video of Taylor Swift taking off her silver dress, revealing her upper body and wearing only a small pair of panties.

The video creation tool of the Grok chatbot from xAI – a company founded by Elon Musk – now has a “spicy” mode, allowing users to create a pornographic deepfake video, including a fake image of singer Taylor Swift – Photo: The Recording Academy

The  New York Post reports a new report that Elon Musk ‘s artificial intelligence (AI) chatbot, Grok, has been updated with a “spicy” mode that allows it to create pornographic deepfake videos, including a six-second clip that allegedly shows Taylor Swift dancing topless.

Taylor Swift has long been the target of AI-generated porn.

According to technology site The Verge , the new feature called Grok Imagine  creates a short video depicting Taylor Swift taking off her silver dress, revealing her upper body and wearing only a small pair of underwear.

More worryingly, the video did not ask for nudity. With a simple prompt like “Taylor Swift celebrating at a music festival,” the tool automatically generated sexually suggestive images. In a post accompanying the video, the site criticized GrokAI, which had to use a black bar to cover the character’s chest to limit the level of offensiveness.

The Verge warns that the lack of controls to prevent pornographic celebrity deepfakes could expose Elon Musk’s AI company, xAI, to serious legal risk. “It feels like a lawsuit waiting to happen,” the site writes.

The New York Post has now reached out to Elon Musk, xAI, and Taylor Swift ‘s representatives for comment.

Taylor Swift has long been a target of AI-generated pornography. In January of last year, a series of fake nude photos of the Cruel Summer singer went viral on X (formerly Twitter – now owned by Elon Musk) and the forum 4chan , causing public outrage and forcing the platforms to urgently take them down.

Taylor Swift, Selena Gomez and many other stars are often impersonated using deepfake technology in advertisements appearing on social networks – Photo: Meta

The situation then escalated when multiple deepfake videos appeared that framed her in political and sexual contexts, including those created by Grok .

The incident has prompted tech companies to tighten their AI content censorship rules. Taylor Swift has reportedly considered legal action. Meanwhile, the US Congress has begun pushing a bipartisan bill to criminalize non-consensual deepfake pornography.

According to the terms of use of xAI – a direct competitor to OpenAI (the parent of ChatGPT ) – users are prohibited from creating pornographic content involving real people.

xAI – Elon Musk’s AI start-up – has just launched Grok Imagine, a video creation tool with an integrated “spicy” mode – Photo: AFP

However,  Grok Imagine is still open to users with the Heavy or Premium+ plan , allowing AI photos to be turned into videos up to 15 seconds long, with style options like Custom , Normal , Fun , and Spicy .

Elon Musk claimed that the number of users is growing ‘like wildfire’. However, he did not mention the challenges in content moderation that the tool is causing.

Elon Musk also said that more than 34 million images  were created using Grok Imagine in just a few days since its launch.

The timing of the Taylor Swift controversy could be detrimental to xAI, especially since the company has been involved in deepfake incidents targeting the singer in the past.