# Grok AI Liability and Section 230 Legal Shifts

**Podcast:** The Journal.
**Published:** 2026-01-27

## Transcript

Elon Musk's AI chatbot Grok is facing intense criticism, accused of allowing ex users to generate fake sexually explicit images.
Late last month, the popular AI chatbot Grok came under fire for a new feature.
An influx of explicit content coming after Grok recently enhanced its image generation abilities with a new model.
Grok, which is integrated on the platform X, began allowing users to edit images with text prompts.
Within days, X was flooded with non-consensual AI-generated images.
Grok is producing thousands of undressed images per hour on X allegedly.
And experts are saying that the scale is nothing like anything they have ever seen before.
Elon Musk, who owns X, has called criticisms of Grok an effort to suppress free speech.
Ashley St.
Clair is a 27-year-old conservative influencer who's known in part for having a child with Musk.
St.
Clair said Grok undressed images of her and depicted her in sexually explicit poses in response to user prompts.
Here she is on CBS.
The worst for me was seeing myself undressed, bent over, and then my toddler's backpack in the background.
Because I had to then see that and see myself violated in that way and such horrific images, and then put that same backpack on my son the next day.
St.
Clair decided to sue XAI, the company behind Grok.
AI should not be allowed to generate and undress children and women.
That's what needs to happen.
St.
Clair's lawsuit gets at the heart of the thorny legal issues surrounding new AI tools.
And it confronts the question who is responsible for the content that users prompt chatbots to create.
She's known for litigating online sexual harm cases and has a particular strategy for holding companies liable for online content.
Carrie told me her goal is to help create new guardrails for an era of artificial intelligence.
It's Tuesday, January 27th.
Coming up on the show, a conversation with the lawyer taking on Grok and XAI.
Carrie Goldberg has built her reputation around internet abuse cases.
When I spoke to her, she was coming to the end of her workday at her Brooklyn law firm, which has a memorable tagline.
Suing, a-holes, psychos, trolls, and pervs, and toxic cases against tech.
Carrie's mission is a personal one.
She started her firm in 2014 after she said that an ex-boyfriend threatened to share intimate photos of her.
At the time, most states, including New York, didn't have laws to protect people from that.
And so I got this unfortunate education, and I started my law firm because I felt that other people needed help against relentless stalkers.
And I quickly started getting cases of back then it was called revenge porn.
And so I was like, how, you know, this is intentional infliction of emotional distress.
How are these companies existing?
We should sue them.
And that's when I came up against Section 230 of the Communications Decency Act.
Section 230 is considered the bedrock of the internet.
It was enacted in 1996, and it protects websites and social media platforms from being held legally liable for the content that users post.
The law was meant to encourage free speech, and it helped the internet take off.
Supporters say that without Section 230, internet discourse as we know it wouldn't exist.
Platforms and sites would much more heavily censor user reviews, comments, and opinions, or simply avoid hosting user content at all for fear of being sued.
But critics say the law has been a way for tech companies and platforms to avoid legal liability for users doing things from selling weapons to posting hate speech and obscene content.
Carrie wanted to hold a company accountable even with Section 230 in place.
She needed a strategy.
And she had an idea.
She came up with a new take on an old legal theory, product liability.
So product liability is an area of law where you're holding companies responsible for the products that they release.
And so companies can be held responsible if they are releasing defective products, defectively designed, defectively manufactured, products where there aren't adequate warnings, they can be held responsible.
Product liability cases have led to things like better airbags and cars and safer beds for babies.
But product liability hadn't typically been used against online platforms until 2017, when Kerry filed a case against the dating app Grinder.
The case involved deep faked profiles of Carrie's client.
Carrie argued that Grindr was designed with quote, foreseeable harm, because at the time she said the app wasn't capable of screening and blocking known dangerous users.
We were like, well, okay, you are a dating app that relies on geolocation technology.
It's an absolute certainty that sometimes your product will be misused by rapists, stalkers, or other kinds of predators.
So if you've not built in into your product technology to ban those abusers, then you've released an unsafe product into the stream of commers.
And so in the case in the Grindr case, the argument you were making was that this product, Grindr, this app, was flawed, was designed in a way that was causing harm and needed to be remedied.
That's precisely the argument.
Grinder fought the case in court using Section 230.
And what happened?
A spokesperson for Grindr said the company is continually evaluating and enhancing its safety measures to keep bad actors off the platform.
But that theory since has been very effective in other cases and other lawsuits.
Like in 2021, when Carrie sued a video chat website called Omagle on behalf of a minor.
In that case, she argued in part that Omagle, as a product, failed to adequately warn child users of adult predators on the site.
The two sides agreed to settle.
Following the suit, Omegle shut down.
When Ashley St.
Clair approached Carrie about suing XAI over Grok earlier this month, Carrie says she saw another opportunity to apply her product liability theory.
This is one of the best arguments I've ever had when it comes to overcoming a tech company's defense of Section 230 immunity.
In the lawsuit, Carrie and St.
Clair alleged that XAI should be liable for its product Grok, because the chatbot is quote, unreasonably dangerous as design.
We are saying that XAI, because of its Grok feature that undresses people, is not a reasonably safe product, and that it was foreseeable through its design and manufacture and its lack of warnings that it would cause injuries like what befell Ashley.
Carrie filed St.
Clair's lawsuit against XAI on January 15th.
Just one day before, X said in a blog post that it had put in new measures to prevent Grok from, quote, editing the images of real people and revealing clothing.
How would that affect your argument if X is saying that it had made these design changes already?
Well, first of all, my client was still unclothed after Grok made that representation.
I have images of her from the 15th of January, where she like new images have been created by Grok of her undressed.
But secondly, we are thrilled that XAI has made that change.
But that doesn't account for the fact that they already cause all these injuries to Ashley and women and children at a mass scale, and they need to be held accountable for that.
The lawsuit was filed in New York and has since been moved to federal court.
It's currently in preliminary stages.
In a court filing, XAI's lawyers said that St.
Clair's claims are subject to dismissal under Section 230 of the Communications Decency Act.
XAI has also launched a countersuit against Ashley St.
Clair in Texas, claiming she's breached the company's terms of service agreement with her suit.
One thing Carrie is banking on with St.
Clair's lawsuit is that chatbots are relatively new.
Because of that, she says they might present an opportunity for the courts to rethink their interpretation of the law.
I had for a long time been, you know, thinking about this idea that XAI, which owns Grok, should not be immune from liability under Section 230, because Grok, not a third party, is the one that is actually generating this material.
These companies are liable for their own content.
But couldn't XAI argue that Grok only creates those images at the request of users?
And so like it's the users, that third party who are liable.
I mean, does that complicate the argument at all?
I mean, I I don't see that.
You know, somebody typing in a prompt is material different from Grok, you know, creating an actual image.
Section 230 is intended for situations where an online platform is just acting as a passive publisher, not where it is itself creating the actual content.
This situation, Grok is not working in the capacity as a publisher, it's actually spitting out the content.
It's generating a content, is what you're saying.
And I mean, certainly you could say that a third-party user is also contributing to the content, but that doesn't mean that Grok isn't.
In early January, Elon Musk said, quote, they just want to suppress free speech in response to criticism of Grok's image generation.
I just want to kind of parse that.
Is there any truth to that statement?
In the sense that there's a risk to free speech when we restrict what people can create with a chat bot like Grok.
Well, I think that argument might be valid if it came to the government trying to create laws that restrict speech.
But when somebody's been harmed in a foreseeable way by content, I don't see that argument flying in a situation where Grok is itself spitting out the content.
After the break, we look at what the law says when it comes to deep fakes.
And Carrie tells me why she thinks the courts are still the best venue for victims to be heard.
Grok isn't the only AI that can declothe images of real people.
But one difference with Grok is that the tool is integrated with a social media platform, which meant the images generated on X went public right away.
So even if XAI, you know, makes good on its promise to stop, you know, producing this content, those images exist forever.
Um, and they're circulated and they're seen by people.
Carrie says the public nature of these images is why she made another legal claim in the lawsuit against XAI, that Grok amounts to a public nuisance.
Public nuisance law addresses things like noise, cleanliness, or safety in public spaces.
Because you know, there were a lot of people harmed in a public space, it allows for us to have, you know, reasonable facts to plead that XAI was acting as a public nuisance, that it was operating in the public sphere and harming lots and lots of people.
And it really lends itself beautifully to this specific product that has long been calling itself the public square of the internet.
The St.
Last year, Congress passed the bipartisan Take It Down Act.
The law makes it illegal for a person to post a non-consensual, sexualized deepfake of someone else.
And starting in May of this year, it also requires that social media companies take down deepfakes at a user's request within 48 hours.
Critics say the law could lead to censorship.
Still, Carrie says laws like the Take It Down Act fall short in important ways.
Well, I mean, I prefer to just go about things using our court system because it's, you know, that makes it so that just me and one client can get a ruling and that can become precedent.
You know, and I want more laws, if any are necessary, that give victims a new cause of action so that they can be in the power seat, and so they don't have to experiment with claims like product liability, and they can actually just use, you know, a deep fake claim that specifically is tailored to this exact behavior.
I think the thing that's so compelling about these digital fraud cases is that it can really happen to anybody.
Like if you have a face, you can you can be deepfaked.
And you know, there's always been a lot of victim blaming when it comes to victims of other kinds of image-based sexual abuse.
Like that person, you know, was stupid enough to take the picture or to send the picture, share it with that unreliable person.
But with deepfakes, everybody listening could become the victim of you know, technology altering your image.
It's interesting to me though that you said, you know, you would rather take this through the courts and set precedent in that way rather than see laws be passed.
Can you explain why that is?
What's the benefit of doing it through the courts versus you know, potentially through Congress or through state law?
Immediacy.
So when I sue, I mean, I, you know, filed my lawsuit within nine days of when Ashley was experiencing the the impact of this.
The thing about regulation and laws is that they're always catching up to the times.
And so I think that they're necessary and they're great, but you know, they respond to things that have, you know, been a problem for, you know, sometimes years.
So I mean, I want laws, but I also want to just be able to sue and go rogue in court.
Ultimately, Carrie, what do you hope will come of Ashley St.
Clair's case against XAI?
Well, I want to get into discovery, and I want to show how you know the quantity of images that were created, the number of other victims that were harmed.
And I want this to set precedence so that this company and its competitors don't go back into the business of peddling in people's nude images.
I want to know what happened in the boardroom.
So, what happened when they like found out that all these people were being harmed?
How much longer did they continue to have this product unleashed on the general public?
So I want to see what was happening on a high level before they actually took action.
What happens if you lose or if the case gets thrown out?
Like, are there consequences?
Could it set back efforts to get compensation for others who've suffered?
No.
I mean, we would appeal an appeal and appeal.
You know, usually when it's when it's a case like this where I know in my guts that it's the right theory, I will keep suing under it until it works.
Carrie, thank you so much for your time.
Thank you, Jess, for having me.
I appreciate it.
That's all for today.
Tuesday, January 27th.
The journal is a co-production of Spotify and the Wall Street Journal.
Additional reporting in this episode by Georgia Wells.
Thanks for listening.
See you tomorrow.