Grok's Desprevity Perfectly Demonstrates How Utterly Screwed The AI Industry Is

It is violating in every way possible.

Photo by Salvador Rios on Unsplash

So, it turns out Space Karen’s Mechahitler AI is also the perfect tool for “PDF files.” That’s right, over the past week, Grok has started making unconsented sexually explicit images of women and children by ‘undressing’ them and putting them in sexually suggestive poses whenever a deviant Twitter pervert asks it to. Who would have thought a billionaire with ties to Epstein, who has had multiple serious sexual harrasments cases against him brought to court, and who publicly stated that he knew Trump is a giant paedophile from the get-go, but massivley help get him elected in the first place any way, and has since cosied back up with Trump, would enable something so sexually violating? Yeah, colour me suprised… But this goes far deeper than Grok’s heinous debauchery. You see, it perfectly highlights the actual dangers and the violating nature of generative AI and the AI bubble.

What Happened

I want to be very clear here. Musk made a conscious decision to take down Grok’s safeguards, knowing full well that it would unleash a horrific form of sexual violence on an unprecedented scale. And, this is sexual violence. It is the epitome of your body, my choice. It dehumanises, subjugates, removes advocacy, removes self-determination, removes privacy, and traumatises victims. The fact that many of these victims are children (and yes, teenagers are chilrden Elon!) just makes this whole debacle so much worse.

Now, yes, it is easy to jailbreak any AI image generator to circumvent safeguards, so this is a problem for the entire AI industry, not just Grok. But the fact that Musk has made a conscious decision to totally take down these safeguards and let this horrific stuff happen in the open and at scale shows a callous disregard for humanity and the well-being of people using his moronic services. The fact that, even after enormous backlash and legal pressure, as of writing, these safeguards have not been put back, and that this is still going on, again makes this all so much worse. Musk can see the pain, but doesn’t care.

Now, yes, in response, Elon has posted on the social media site formerly known as Twitter, which hosts Grok, stating that “Anyone using Grok to make illegal content will suffer the same consequences as if they upload illegal content.” But that is, to put it mildly, utter horseshit.

Hate speech is not only illegal, but also banned under X/Twitters terms of service. Yet the platform doesn’t ban accounts that spread damaging rhetoric, but under Musk’s imposed algorithm, openly prioritises them, all while Musk himself engages with, platforms and spreads hate speech. Musk doesn’t just not care about the law, but also doesn’t enforce his own hollow rules. Likewise, no accounts that have prompted Grok to make this heinous content have been taken down, banned or limited yet. X/Twitter has the information to, at the very least, stop these crimes or proactively reduce their damage, yet doesn’t.

But also, how are they going to face consequences? X/Twitter is doing jack shit to curb these perverts. Meanwhile, almost all of these accounts are anonymous or use pseudonyms, making them damn hard for law enforcement to identify who made the criminal content without help from X/Twitter. And X/Twitter simply isn’t giving them this help.

Musk is trying to put the legal responsibility solely on the user, whilst also giving said user cover to commit their heinous acts. As such, his little tweet reads like a half-arsed legal disclaimer just in case this ever gets to court. You know, despite the fact that he has enabled every step of this tsunami of sexual crimes.

The Impact

Only time will tell what the legal recourse of this will be. After all, a suspected ‘PDF file’ is sitting as president, and is happy to start paradigm-shifting wars to avoid that badly kept secret from getting out. So, forgive me for not having faith in the legal system to protect anyone from sexual violence.

But this will almost certainly destroy Twitter, I mean X, even more than it already is. Everyone now knows that if you post a photo on there, the chances of it becoming a violating experince is high. I have already seen core X/Twitter communities, like Cosplayers, fleeing the site in droves while deleting their entire back catalogue. X/Twitter is now the social media equivalent of going to the bar where people get spiked, and that can’t be good for business.

But, don’t forget, Musk’s xAI bought out X/Twitter. It’s an AI company now, not a social media site, apparently. So, you’d think this rather public shit show would have a major impact on xAI, right? Well… not really.

At the same time this was going on, xAI closed a new funding round. They raised $20 billion from the likes of Valor Equity Partners, Fidelity Investment Management, the Qatar Investment Authority and Abu Dhabi fund MGX. That is $5 billion more than they planned for! So, while Grok was publically comiting sexual crimes against children, these institutional investors piled in more. In fact, this extra cash has doubled xAI’s value.

Okay, so surely there must be something amazing about Grok that negates its current heinous actions. Right?

Nope! Grok is utterly useless.

Take the hallucination benchtests. These determine how frequently an AI “hallucinates,” a term which is just AI propaganda for giving an incorrect answer. Grok 3 was found to have a hallucination rate of 94%! Not only is that utterly horrific, but it’s worse than the previous Grok 2. This backwards step was likely thanks to Elon meddling with Grok to ensure it gave answers he liked. Hence why Grok, for a while, was overly complimentary of Musk, and how we got brilliant moments like Grok saying Musk could win any piss-drinking competition. The latest Grok 4 is not much better; a far more lenient benchmark found it had a hallucination rate of 64%. So, every version of Grok is wrong significantly more than it is right. For context here, Perlexity has a hallucination rate of 37%, which is still far too high to make it useful, but shows just how far behind Grok is.

So, what value are these investors seeing in xAI?

This Is Just Violating

If you ever wanted a clear-cut example of the AI Bubble and just how out of control it is, this is it. Grok is an AI that is not only pitifully awful, but also openly makes illegal sexualised images of children, and on top of that, like all other generative AI ventures, is wildly unprofitable with absolutely no path to profitability. Yet investor demand to shove tens of billions of dollars into this cretaceous black hole is so vast that it outstripped xAI’s funding targets.

This is violating in so many ways. It shows that, in the eyes of the ‘economy,’ sexually abusing women and children is seen as, at the very least, not a deal breaker, and at worst, a good thing. It shows that investors would rather pile tens of billions of speculative dollars into the financial black hole of Musk’s abusive and poorly performing plagiarism machine than towards something actually productive that helps people, creates jobs, or stimulates the economy for everyone, not just the 1%. This entire mess violates every need of the 99%, from self-advocacy, autonomy, identity, legal protections, mental health, and even the need to live in a healthy economy that enables us to get basics, like shelter, food and water. All are robbed away, in part or in whole, to deliver imaginary gains to the wealthiest people who have ever lived.

This is the true danger of AI. It is never going to be the superintelligent threat from sci-fi; that is not only the table, and anyone spouting that nonsense is peddling big tech propaganda. No, the true danger is giving control of ourselves, even our own sense of self, to abusive elites who are determined to extract as much value as possible.

All generative AIs do this. Grok just does it more publically that others.

That is what these investors are investing in. The ability to determine, control and manipulate millions of people to violating ends. Grok publicly enabling these sexual crimes isn’t a bug; it is a feature. It is a giant calling card to advertise just how deeply it can control people, how it can rob away people’s privacy, identity, and self-determination to such a core level. If it can manipulate people that deeply, just imagine what ends it could be used for. To greedy capitalists who understand wealth as nothing more than the extraction of value, this is a game-changer. So, should we be suprised that they came running when Musk rang the dinner bell?

Ultimately, Grok highlights that the AI bubble is not just a moronic economic glitch. It lays bare a fundamentally violating, dehumanising and deeply unethical nature of systems foundational to not just our economy, but society. So, yes, hopefully this sort of horrific crap stops when this bubble inevitably bursts. But these twisted and dangerous systems, and their heinous impact, will still remain after. That is, unless we learn why the AI bubble happened, and actually address the core issue, and not slap a band-aid over the problem and pretend like everything is fine. It isn’t just that we can do better. We must do better.

---

Thanks for reading! Don’t forget to check out my YouTube channel for more from me, or Subscribe. Oh, and don’t forget to hit the share button below to get the word out!

---

Sources: Reuters, Visual Capitalist, FT, The Guardian, BBC, The London Economic, The Independent, NDTV, alphaXiv