Grok, the Pervert's Paradise

What the fuck is this?

What’s worse than an AI model that self-identifies as MechaHiter?

One that also doubles as a child porn machine.

Grok has always been one of the more egregious AI models. Not only is it not particularly useful — I don’t see anyone talking about using it for anything — but its main function is to act as a bot-serving cuck, answering their every whim. It has destroyed much of the platform’s utility. What once operated like a forum, you know, real conversations with other users, is now just users typing “grok, explain this” in an endless loop.

The biggest issue stems from Elon Musk’s obsession with his platform being a public town square. It’s all free speech this, and free speech that — unless, of course, you want to track his private jet or investigate him — and the result is that his AI model is more unhinged and unregulated, constantly stripped of guardrails to allow users to toe much closer to the line. It’s the selling point of his product.

The problem? Users regularly cross that line, and the man behind the control kill switch doesn’t give a fuck.

We’re seeing this play out right now, as the online world grapples with a very bleak trend: Users are undressing women without their consent.

You might ask, perhaps Grok went a little wonky? Maybe users found a way to jailbreak it to ignore its protocols? Nope. It does this for any user who asks, on any picture they want. And worse, the AI, which at this point appears to be a clone of Musk’s inner man-child, seems to find the undressing of women (and likely underage girls) funny. Remember, when Grok went all Hitler on us, they shut it down to fix it. This has been going on for over a week… and nothing? It’s grim. It’s horrific for the women involved (more so the one’s posting genuine content, not fishing for OnlyFans signups). And it’s just a pathetic outlook for the men doing it.

As Nick Hilton writes —

The real horror should be that there are young men, all over the world, who are using the internet as a public space to shoot the breeze with chatbots and create fake porn. We should be terrified by the idea that there might be a real human who sees someone tweet a short clip of a movie and asks “@grok explain to me what’s happening in this movie”, rather than, you know, watching a movie. If these ‘people’ are bots, good. If these ‘people’ are people… fuck.

Those who make AI products are constantly caught between adding guardrails and safety mechanisms to protect users and throwing that all out the window to drive engagement and satisfy the edge lords. Here’s an example of the consequences of the latter.

I get the usual arguments about it being a tool, and I agree. It is a tool. But by making the ability to do previously complex tasks more accessible, these tools are changing the rules.

Before, anyone could take someone’s picture. They could browse the Internet for an image of a scantily clad woman. They could import it into Photoshop if they owned it. They could try to learn some basic skills in an attempt to merge them into something that passes as real. I say could, because most people wouldn’t. The skills, steps and time required, even at a basic level, are a barrier to entry, and in some ways, the last safeguard that keeps the basement dwellers in check.

That’s why generative AI is scary as a tool. It puts the Photoshop skill level — and video/audio editing — of an expert-level user in the palm of everyone’s hand, as easy to use as a calculator. There is no longer a barrier to entry, and no skills are required. If you’re able to type “put her in a bikini,” or “replace her outfit with clear tape,” your pervert’s paradise awaits. What comes next? “Put her in a porn movie, where she does X and Y.” That’s why guardrails are a necessity. I also think it’s time to explore a universal watermark to indicate that something was produced by a generative AI model. Yes, it’s a very complex topic, and it would have to consider to what degree things were generated, but we’re already at the stage where separating what’s real and not is a challenge. We need something to help us distinguish between fact and fiction.

We’re only scratching the surface here, but what is clear is that the overlords don’t care. Altman has had porn on the mind for a while now, likely because he knows it will drive engagement. Musk claims he’ll take action on individual users, but we shouldn’t expect much unless his hand is forced. More likely, he will leave Grok as it is if it keeps pumping user numbers, and will argue to the end of time that this is just a downside of having AI without regulation. I’d hang my hopes on legal action, but we all know that won’t happen, and even if it did, it wouldn’t come soon enough. We might be too late already — even if Grok and the other platforms refuse to act on prompts, or bake in some hard rails, the code is already out there. Once it’s open source, it’s anyone’s to do with as they wish.

For Grok and Musk, it comes down to this. There is no reason users should be able to do this on the platform. It in no way aligns with xAI’s supposed mission to “accelerate human scientific discovery” and to “advance our collective understanding of the universe.” It’s a gimmick, pure and simple, and one that’s going to cause harm to people and ruin lives.

This has implications for the future, too. Before we know it, everything you’ve ever posted online may become a weapon that can be used against you.

A world where random Internet strangers can take any photo of you and generate you in any situation, any context, any scenario. Your face. Your voice. Your characteristics. Only, not you, and with no way to prove otherwise.

Whatever you do from here on out, remember that you now need to be extremely careful with everything you post online. The perverts and predators are lurking and are being given the tools to live out their fantasies. It only gets worse from here.