Non-consensual intimate imagery harms real people.

CSAM normalizes and facilitates abuse of real children.

Grok, and everyone involved in it or similar endeavours, facilitate abuse.

In a way, leaving it open as a honeypot is the best action.

If you build a system explicitly designed to have no content boundaries, and it produces CSAM, that's not a failure of execution - that's the system working as designed. You don't get credit for noble intentions when the outcome was entirely foreseeable.

Deciding to place no limits on what an AI will generate is itself a value judgment. It's choosing to enable every possible use, including the worst ones. That's not principled neutrality; it's moral abdication dressed up as libertarianism.

When I was young it was considered improper to scrape too much data from the web. We would even set delays between requests, so as not to be rude internet citizens.

Now, it's considered noble to scrape all the world's data as fast as possible, without permission, and without any thought as to the legality the material, and then feed that data into your machine (without which the machine could not function) and use it to enrich yourself, while removing our ability to trust that an image was created by a human in some way (an ability that we have all possessed for hundreds of thousands of years -- from cave painting to creative coding -- and which has now been permanently and irrevocably destroyed).

Are those goals noble? This is the same guy who also said "with AI we are summoning the demon" and whose self-justification for getting a trillion dollar Tesla bonus deal involved the phrase "robot army"?

Fortunately this narration did not catch traction.

I'd assume they were just blindsided by the response; they're likely in real danger of getting either DNS-delisted or outright banned in several jurisdictions.

Bloody hell, what the hell is wrong with people?

2) Inability to differentiate between real and fake - less theoretical.

EDIT: op seems to have clarified in another reply that they are talking about fully generated computer images but this is out of context, the outcry here is that Grok generated fake images of actual people to start with!

Huh? People have been editing images of real women to depict them in bikinis (and that's the least offensive).

There is also a greater debate about giving people with harmful deviances a “victimless” outlet versus normalising their deviance. Sure there are no children harmed in making the images, but what if they generate images of child actors, or normalise conversation about child sexual abuse?

People are being hurt by this, because "just pixels on a screen and bytes on a disk" can constitute harm due to the social function that information serves.

It's like calling hurling insults at someone as "just words" because no physical violence has occurred yet. The words themselves absolutely can be harm because of the effect they have, and also create an environment that leads to further, physical violence. Anyone who has experienced even mild bullying can attest to that.

Furthermore women and girls are often subject to online harassment and humiliation. This is of course part of that -- we aren't talking about fictional images here, we are talking about photos of real people, many who are children, being manipulated to shame, humiliate, and harass them sexually, targetted at women and girls overwhelmingly.

Advocating for the freedom to commit that kind of harm against other people is gross, and you should reconsider your views and how much care you have for other people.

However, I think it's clear that Pandora's box is now wide open and that you cannot close it. Sure, you can turn off that Grok integration, but the AI image generation capabilities are now widely available for basically anyone to use.

I wonder whether it'd be better to just "accept and live with it". I agree that this can cause a lot of harm, but I don't see a way where this can be outlawed and prosecuted in such a way that there's a net benefit for society. In the EU many have been battling proposals like Chat Control for the past decades not because they want to protect sex offenders, but because backdooring societies privacy on a grand scale is likely far more detrimental than the impact of sex offenders. (And here we aren't even talking about "real" CSAM content.)

The level of discourse on twitter is pretty terrible already but at that point what’s the platform even selling… and who is the platform for?

Just from a product standpoint you got problems.

I don’t think we can avoid a world where people can generate CSAM easily, so we have to separate the discussion between being able to do that privately and grok being able to do it.

It makes sense to me that we don’t want widely used websites to contain images of CSAM that you can’t easily avoid, it’s simply repulsive to almost everyone and that’s almost certainly a human instinct, I don’t think it needs to be much more complicated than that.

In terms of generating CSAM privately or even sharing it with other people, I think this is a much more interesting discussion. I think at this point it is an open question on whether it is harmful. Could this replace the abuse that is happening to create some of the real content? Does the escalation argument hold water - will people be more likely to sexually assault children due to access to this material? I don’t think we know enough about pedophilia to answer these questions but given that I don’t think there is a way to stop generating this content in 2026 we really need to answer these questions before we decide to simply incarcerate everyone doing it.

So with that in mind, why should we assume that the unrestricted proliferation of fake child porn would not produce significant harm downstream to society? Is the assumption that unlimited fake child porn would satiate the kind of people who now seek out real child porn?

I would strongly dispute that as the amount of energy required by and pollution produced by AI systems does cause a lot of harm to living creatures and environments.

Is it because it was pre-trained with real images? (which would be highly illegal and immoral, but I wonder if Twitter has a data-curation team somewhere)

Maybe some kind of distillation technique such as "generate normal porn -> decrease body size -> generate childish face and replace the original image's face with it"? Which would prove there's an intent to explicitly allow generation of this kind

Is it an emergent behaviour?

Why are there no better safeguards?

Then you are not talking about the article, so what was the point of your comment?