For reasons known only to God and Tim Sweeney, the Epic Games CEO felt compelled to offer his two cents on the current issues with Twitter's AI bot, Grok. Grok being the generative AI that is still modifying and creating sexually explicit images of others, mainly women and children. Many of these created images by Grok sexually depict women and children being "sexually abused, humiliated, hurt and even killed."
Twitter and Grok's developers at xAI did the absolute bare minimum in response to the creation of child sexual abuse material (CSAM). They restricted image creation to only paid subscribers, a "solution" that most find to be pathetically weak. The powers that be behind Grok have the ability to stop the AI from creating the sexual material but have instead simply tucked it away behind a paywall that everyone can still access.
Researchers such as Genevieve Oh found that "the chatbot generated about 6,700 (images) every hour that were identified as sexually suggestive or nudifying." According to Oh's research, the other top five websites that produce similar content were averaging just 79 images per hour. Oh's research took place over a 24-hour period from January 5 to January 6.
The fact that Grok is still producing those sexually explicit images of women and children pushed several countries and lawmakers to consider blocking access to Twitter entirely. At present, both Indonesia and Malaysia have banned Grok. A UK watchdog, Ofcom is presently investigating Elon Musk and Twitter over Grok's image creation. If the law was found to have been broken, they could fine Twitter up to 10% of its worldwide revenues or £18 million, whichever is greater. The UK is also not ruling out an outright ban on Twitter.
In the United States, at least three Democratic senators are calling on Apple and Google to pull both the Grok AI and Twitter apps from their mobile storefronts. Senators Ben Ray Luján, Ed Markey, and Ron Wyden wrote an open letter to Apple CEO Tim Cook and Google CEO Sundar Pichai to "remove the (Twitter) and Grok apps from their app stores until the company's Chief Executive Officer, Elon Musk, addresses these disturbing and likely illegal activities."
Over the weekend, Epic Games' Tim Sweeney decided to weigh in with his own thoughts. In response to a news article talking about the US lawmakers' open letter to Apple and Google, the Epic Games CEO accused lawmakers of only wanting to ban Twitter so that they can "censor all of their political opponents."
All major AIs have documented instances of going off the rails; all major AI companies make their best efforts to combat this; none are perfect. Politicians demanding gatekeepers selectively crush the one that's their political opponent's company is basic crony capitalism.
One Twitter user asked, "Tim do you realize what you're defending"? To which Sweeney replied:
I defend open platforms, free speech, and consistent application of the rule of law. The bad stuff people do with AI, I do not defend, but I staunchly oppose the wrongdoing of a few from being used as a pretense to undermine the freedoms of all."
This is a vile lie by PC Gamer. I criticized a government official for pressuring Apple and Google to block a speech app owned by their political opponent, deplatforming 500,000,000 users on the pretense of stopping a small number of users from distributing disgusting content.
Sweeney also dove into the conspiracy theory hole when responding to a Twitter user that said they appreciated Sweeney speaking up in the face of what they see as "a power play" by some nebulous boogeyman.
It’s almost as if there is some group chat somewhere that is used to plan who to attack with this ludicrous “if you don’t support banning X you must be into bikini pics of Keir Starmer” narrative.
Were this a political attack on Twitter, there have been plenty of instances in the past where lawmakers could have pushed for the app to be removed from Apple and Google devices. Since Musk took over Twitter, the site has run rampant with unchecked bigotry, racism and white supremacy, abuse, credible threats, misogyny, and intolerance. Whatever you may have thought of Twitter's lax moderation pre-Musk has only gotten far worse in recent years.
It is my personal opinion that Tim Sweeney and people like him, those that are going to bat for a platform that allows users to generate CSAM, should just shut the actual fuck up. If your idea of "free speech" means allowing platforms to decide if any amount of CSAM can be created or shared by their own AI bot, then I say fuck your free speech. It costs you literally nothing to not jump to the defense of a multi-billion dollar company that is doing nothing to stop their AI from creating sexually explicit images of women and minors.
It should not matter if it's just "a small number of users" creating CSAM. It's the fact that it's allowed to be created at all that is the issue. It's the fact that Twitter and xAI have done fuck all, save for a pitiful paywall, to stop the creation of that content that is the issue.
Take a moment to look at the situation and realize exactly what it is you are so staunchly defending.
I also have to assume they don't just want to simply solve the problem by shutting down Grok for the same reason no one else has shut their AI technology down due to this, because it's the fastest growing most out of control piece of technology the world has ever seen, AI is being integrated into every facet of society including government, i doubt the UK government wants anything to do with Grok considering their relationship with Elon so they wouldn't mind banning it and diverting custom toward the AI systems they are using instead, and if they can take down X at the same time it's a bonus as i do (painfully) agree with Tim it is an extremely disruptive part of modern politics, i think most agree X helped deliver Trump to the White House as the most obvious example.
I think locking it behind premium was the only thing they could do until they have a solution that isn't shutting it down, i know it sounds like i'm defending them but i can't blame one AI generator just based on popularity if other services are doing it too, they should all be banned, but i'm just going by what others have said about that as i have no idea myself.
My random guess as to why governments haven't already banned images of children being anywhere close to AI is for its use as training data in identification, CCTV, age and facial recognition etc... These are systems governments would love to automate in their ever growing desires for control, they need AI to be able to achieve anything, which is why i think AI related legislation has been virtually absent.
So no, putting it behind the paywall and still allowing the creation of literal illegal material is not the only thing they could do. It's not hard to stop unless you literally try nothing, which is exactly what xAI has done: Nothing.
They could even disable the creation of explicit materials if they want to keep their image generation running. Tons of other AI operators have been able to do that no problem when faced with related issues. Same as when AIs are caught creating content based off of copyrighted content. Plenty of AI systems had to quickly resolve that. Most of the mainstream AIs do NOT do allow for the creation of any sort of explicit content, period. There are AIs that are purpose built for creating sexual imagery but NOT CSAM! Here you have a mainstream AI with zero guardrails and run by one of the world's biggest megalomaniacs. Need it be reminded that Grok previously praised Hitler, called itself MechaHitler, celebrated the deaths of children from flooding in Texas, was altered by *someone* within xAI to praise Elon Musk in nearly every response for a time, etc.?
It's not hard to change variables, logic, etc. for these AI.
If you ever had an AI response glitch out to show you the "behind the scenes" logic, you'd know that it's usually as simple as writing literal lines of text to the AI not to do this, not to do that, not to use content that is copyrighted or illegal as the basis for the output. Twitter, Musk, and xAI have done fuck all while Sweeney is making baseless assumptions and *apparently* defending a tool and website that is still creating explicit images of children.
Fuck. That.
I'm just being the devils advocate and suggesting something more complicated is going on, otherwise yes as you say they should just disable NSFW and this problem is already solved.