Grok's 'spicy mode' making pictures with kids is horrifying | Opinion

Grok, the artificial intelligence chatbot connected to X, has been making headlines for the recent “undressing” epidemic, in which users generate manipulated images that feature people in very little clothing and suggestive positions.

Countless people – mostly women, both famous and private – have been affected. The story has gained traction as people began raising flags about many of these images featuring children.

Ashley St. Clair, the mother of one of X CEO Elon Musk’s children, was one of the women targeted by this act of sexual violence. One photo of St. Clair edited by Grok is from when she was 14 years old, the conservative influencer says.

“I really don’t care if people want to call me ‘scorned’," St. Clair posted to X, "this is objectively horrifying.”

She's right. No matter where you fall on the political spectrum or how far you think the First Amendment should go, it’s terrifying that explicit content can be created of children ‒ or any nonconsenting party for that matter.

The problem itself has multiple means of mediation: It’s up to Musk to regulate X, but the private sector and lawmakers must put pressure on the CEO to do something. And while changing the culture is a comically gargantuan task, we must start interrogating the misogyny that begets illicit content like this.

'Undressing' AI isn't a new problem

This has been going on for months. Back in August, The Verge published an article about the deepfakes of Taylor Swift that were being created by Grok to show the singer in scant clothing. Only recently has the problem caught national attention.

Other AI chatbots, like ChatGPT and Google Gemini, are also being used to generate images of women in bikinis, but Grok tends to create more graphic content than the other two. That’s because Grok, unlike other AI chatbots, offers "spicy mode," which allows users to create suggestive content.

You’d think that, as CEO and someone with a connection to one of the victims of this abuse, Musk would take a firm stance against these images. He appeared to comment on the issue on Jan. 3, posting to X that “anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content.”

But while Grok – a nonsentient being – has “apologized” for these images, it doesn’t seem that any real human has taken accountability. Musk has mostly responded with laugh-cry emojis. News outlets that have tried questioning xAI, the company behind X, about the issue have been met with the auto-response “Legacy Media Lies.”

Elon Musk may not care about victims, but advertisers and lawmakers can make him

If Musk won’t listen to victims, perhaps he’ll listen to advertisers. Companies like Google, Apple and Amazon still advertise on the platform in spite of previous controversies. Do they really want their products to appear alongside suggestive images of minors?

Elected officials have tried to stop these deepfakes from occurring. The Take It Down Act, which received bipartisan support in Congress and was signed into law by President Donald Trump in May, makes it illegal to share nonconsensual illicit videos and photos, including deepfakes.

Many states have enacted laws against revenge porn, and there are national laws prohibiting child sexual abuse material. Still, more can be done.

Getting rid of these images currently requires people to report these posts to X. The problem is that they are allowed to be created in the first place. It may even be worth the U.S. Supreme Court taking this up as an issue.

No matter if it’s the private or public sector intervening on behalf of the victims, it’s clear that this cannot continue.

Follow USA TODAY columnist Sara Pequeño on Bluesky:  @sarapequeno.bsky.social