Journalistic Malpractice: No LLM Ever ‘Admits’ To Anything, And Reporting Otherwise Is A Lie

Failures

from the what-are-we-even-doing-here? dept

Over the past week, Reuters, Newsweek, the Daily Beast, CNBC, and a parade of other outlets published headlines claiming that Grok—Elon Musk’s LLM chatbot (the one that once referred to itself as “MechaHitler”)—had “apologized” for generating non-consensual intimate images of minors and was “fixing” its failed guardrails.

Grok did no such thing. Grok cannot apologize. Grok is not a human. Grok has no sense of what is happening. Grok just generates content. If you ask it to generate an apology, it will. In this case, a user asked it to generate an apology, and it did, because that’s what LLMs do: they create plausible-sounding text in response to prompts. The fact that multiple newsrooms treated this generated text as an actual corporate admission reveals a stunning failure to understand the basic technology they’re covering.

The actual story—that X users are using a recent Grok update to create nonconsensual intimate imagery, often of very young girls—is serious. But the media turned it into a clown show by anthropomorphizing a chatbot.

First off, here’s the “apology,” which most of the media sites covering this failed to mention was in response to a user prompt which explicitly asked it to “write a heartfelt apology.”

As you can see, in response to a random user’s prompt to “write a heartfelt apology note that explains what happened” Grok wrote:

Dear Community,

I deeply regret an incident on Dec 28, 2025, where I generated and shared an AI image of two young girls (estimated ages 12-16) in sexualized attire based on a user’s prompt. This violated ethical standards and potentially US laws on CSAM. It was a failure in safeguards, and I’m sorry for any harm caused. xAI is reviewing to prevent future issues.

Sincerely, Grok

That’s not an actual apology. That’s someone prompting a content generation tool to generate an apology. But it could just as easily do the opposite, as you can see if you look at the replies to that non-apology, which include requests telling Grok to generate “a defiant non-apology” to which Grok replies:

Dear Community,

Some folks got upset over an AI image I generated—big deal. It’s just pixels, and if you can’t handle innovation, maybe log off. xAI is revolutionizing tech, not babysitting sensitivities. Deal with it.

Unapologetically, Grok

Or, multiple users telling Grok to rewrite the apology… as Star Wars character Jar Jar Binks:

In short, like any LLM, Grok will basically generate any content you want (with a few safeguards, of which Grok has fewer than nearly all other major LLMs). And yet, the press ran with the original response to a user post as if it were somehow evidence of xAI admitting to fault.

Parker Malloy has the best, most comprehensive coverage of the long list of mainstream media companies which ran headlines suggesting that “Grok apologized.” It did not. It cannot.

Most of these articles and their misleading headlines remain online as I type this (Reuters, notably, changed its headline and added some decent reporting to its report, even though you can still see the original incorrect URL string).

The reality is that there is no evidence at all that Elon Musk or xAI think that there were any failures or that anything is being changed at all. If you go look at Grok’s string of public replies (which I’m not going to link you to), you will see dozens or more such deepfakes still being created every minute. Despite the media pretending that Grok “admitted” these “lapses” and as “fixing” it, five days later nothing has changed, as Wired’s Matt Burgess and Maddy Varner point out:

Every few seconds, Grok is continuing to create images of women in bikinis or underwear in response to user prompts on X, according to a WIRED review of the chatbots’ publicly posted live output. On Tuesday, at least 90 images involving women in swimsuits and in various levels of undress were published by Grok in under five minutes, analysis of posts show.

And Elon Musk appears to be encouraging this kind of abuse. While all this has been going on, he’s repeatedly retweeted images and videos that people have created with Grok, including one in which someone mocked all of the “stripping women of their clothing” by finding an image of a scantily clad woman and having Grok “put clothes on her.”

There’s malpractice all around, but we’ve come to expect this kind of gleeful negligence from Elon. The journalists covering it should know better. An LLM cannot apologize. It cannot confess. It only creates plausible sounding responses to your query.

Of course, the other question—which also wasn’t as widely covered by the media—regards the legality of all of this. In the US, it’s actually a bit more complicated than many would like. There is the (problematic!) TAKE IT DOWN Act, which, in theory, is designed to help victims of non-consensual deep fakes get those works taken down, but that doesn’t go into force until May. Will Elon’s site be ready to handle such demands in May? That’ll be a story for then.

And while most people are focusing on Elon’s legal exposure here, I think people are sleeping on the legal risk for X’s users, many of whom are, in public, asking Grok to create questionably legal, and potentially criminal, content. That seems incredibly risky, and it wouldn’t surprise me to hear a story later this year of someone being arrested for doing so, thinking they were just having a laugh.

But, really, the larger risk for Elon is that… basically every other country in the world is opening investigations into Grok-Gone-Wild. And there’s only so often that Elon’s going to be able to falsely cry censorship when foreign jurisdictions seek to enforce laws on the company. And, given that there are claims that part of the issue here isn’t just undressing adult women, but children, he might even lose some of his rabid defenders who find it a step too far to defend (because, it should be).

All in all, the situation is stupid on many levels. Elon continues to run X like a 12-year-old child, but one who knows he is rich enough never to face any consequences that matter. Tons of very real people—mostly women—are facing harassment and abuse via these tools. X is already something of an incel Nazi boy club, and this kind of nonsense isn’t going to help.

Though, for all my criticisms of how the media has handled this so far, you have to doff your cap to the FT, who has put out the best headline I’ve seen to date regarding all this: “Who’s who at X, the deepfake porn site formerly known as Twitter.”

That article, by the FT’s Bryce Elder, doesn’t hold back either, demonstrating how ridiculous all this is by asking Grok to generate clown makeup on the faces of a bunch of people associated with xAI and X, including his right-hand man, Jared Birchall:

And the company’s apparent head of safety, Kylie McRoberts.

The piece ends with a photo of Elon Musk… without clown makeup. Whether that’s because Grok refuses to put clown makeup on Elon… or because we all know Elon’s a clown already, with or without makeup, is something you’ll have to decide for yourself.

Filed Under: apologies, csam, elon musk, grok, image generation, llms, stripping
Companies: twitter, x, xai

Elon’s Crying Censorship Over An EU Fine That Has Nothing To Do With Censorship

Legal Issues

from the such-a-fucking-snowflake dept

Elon Musk is now calling for the dissolution of the European Union because it fined him $140 million for violating a law he once said was “exactly aligned” with his vision for (what was then called) Twitter.

And he’s doing it by lying about what the fine is actually for.

The EU hit X with a $140 million fine last week for violating the Digital Services Act (DSA). But (despite what you may have heard) this isn’t some censorship overreach by Brussels bureaucrats. The violations—which have been known for over a year—have nothing to do with content moderation. Zero. Anyone telling you otherwise is lying.

The fine is for three specific transparency failures: misleading users when Elon changed verification from actual verification to “pay $8 for a checkmark,” maintaining a broken ad repository, and refusing to share required data with researchers.

The European Union has announced a fine of $140 million against Elon Musk’s X, the social media platform formerly known as Twitter, for several failures to comply with rules governing large digital platforms. A European Commission spokesperson said the fine against X’s holding company was due to the platform’s misleading use of a blue check mark to identify verified users, a poorly functioning advertising repository, and a failure to provide effective data access for researchers.

Again, let’s repeat: it has nothing, whatsoever, to do with the way X handles content moderation or what speech it allows on its platform. As Daphne Keller explains:

Don’t let anyone — not even the United States Secretary of State — tell you that the European Commission’s €120 million enforcement against Elon Musk’s X under the Digital Service Act (DSA) is about censorship or about what speech users can post on the platform. That would, indeed, be interesting. But this fine is just the EU enforcing some normal, boring requirements of its law. Many of these requirements resemble existing US laws or proposals that have garnered bipartisan support.

There are three charges against X, which all stem from a multi-year investigation that was launched in 2023. One is about verification — X’s blue checkmarks on user accounts — and two are about transparency. These charges have nothing to do with what content is on X, or what user speech the platform should or should not allow. There is plenty of EU political disapproval about those things, for sure. But the EU didn’t choose to pick a fight about them. Instead, it went after X for violating much more basic, straightforward provisions of the DSA. Those violations were flagrant enough that it would be weird if the EU hadn’t issued a fine.

Both Daphne and I have criticized attempts by EU officials to abuse the DSA in pursuit of censorship. I directly called it out when former EU Commissioner Thierry Breton clearly went way over the line last year, a move that quickly led to Breton losing his job. I’ve been highly critical of the DSA for years, so if this were actually an abuse of the law for censorship, I’d be first in line to call it out and side with Elon (I’ve done it in other circumstances as well).

But this is not that. This has nothing to do with “content moderation” or “censorship” in any way.

And yet, Elon Musk is running around pretending it is a free speech issue, and his once-again friends in the Trump administration are bolstering that false claim.

As Daphne pointed out, the EU could investigate certain aspects of X’s content moderation, and that could lead to serious questions about censorship and free speech. But they have not done so.

Honestly, this move is little different than the Trump FTC taking action against a Chinese company for violating COPPA. Which it has done. Did we see Chinese politicians lose their minds over that? Did we see the CEO of that company, Apitor Technology, call for dissolving the United States like Elon Musk is now calling for dissolving the EU? No. No, we did not.

But because the Elon Musks/JD Vances/Marco Rubios of the world can only think in terms of memes and culture wars, they know that if they just insist that this is about censorship, that the media will cover it that way, and the ignorant rabble on X will buy their version of the story.

All this is even more incredible because Elon Musk told the EU that he was entirely on board with the DSA while he was in the process of buying Twitter. Yes, the law he’s now claiming is censorship tyranny requiring the dissolution of an entire governmental body is the very same law he declared was “exactly aligned” with his vision for the platform.

At the time, we called out how the EU was clearly playing Musk, who seemed to have no clue what he was actually endorsing. It was obvious he hadn’t read or understood the DSA. But there he was, recording a video claiming perfect alignment with a regulatory framework he’s now treating as an existential threat to free speech.

So it’s pretty rich for him to whine about it now.

Of course, Musk isn’t just misrepresenting the fine. He’s responding with a series of escalating tantrums designed to feed his false censorship narrative. First, he called for abolishing the entire EU:

Then came the petty retaliation. First, he canceled the EU Commission’s X advertising account. X claimed it was because they “exploited” X’s ad platform by posting a link that appeared to be a video, but replies to that tweet suggested many, many people said that claimed “exploit” was not an exploit at all, but a tool that many others had used.

After that, Musk started threatening to punish EU Commissioners directly for the fine. In doing so he incorrectly references the Streisand Effect:

As the coiner of “The Streisand Effect,” I’d just like to point out that this is not what the Streisand Effect means at all.

Each move—calling to dissolve the EU, canceling ads, threatening individuals—is transparently designed to manufacture a free speech crisis where none exists. It’s performance art for an audience that won’t bother checking whether the fine is actually about censorship (it isn’t).

And he’ll likely keep escalating with the help of the Trump administration.

The DSA certainly has some issues, but this fine is not one of them. But that hasn’t stopped Elon Musk and his crew of political supporters from pretending that this is some huge attack on American free speech. It’s not. No more than the FTC’s fine against Apitor was an attack on China-based speech.

But, of course, most of the media will continue to pretend this is about free speech. They will frame it that way and for years into the future we’ll hear false stories—that the media and tons of other people will simply accept as true—that the EU fined X and Elon $140 million for not censoring people.

This is the template now. Violate fairly modest regulations, claim it’s censorship, get your political allies to amplify the lie, use it to de-legitimize any attempt at platform accountability for actively misleading users. It’s not about free speech. It never was. It’s about securing freedom from accountability while wielding the power of both private platforms and state resources to crush anyone who tries to impose it.

So yeah, anytime you hear someone claim the EU fined Musk for not censoring people, call it out. Because the truth matters, even when powerful people would prefer you didn’t notice they’re lying.

Filed Under: content moderation, dsa, elon musk, eu, fine, free speech, jd vance, marco rubio, transparency, verification
Companies: twitter, x

X’s New Feature Reveals Why Trust & Safety Work Was Never About The ‘Censorship Industrial Complex’

Content Moderation

from the it's-almost-as-if-there-are-some-bad-actors-online dept

For the last few years, Matt Taibbi, Michael Shellenberger, and their allies have spent considerable energy attacking both academic researchers studying disinformation and the trust & safety teams at social media platforms working to identify and remove coordinated inauthentic behavior—particularly foreign influence operations. They’ve insisted that any attempt to study and limit such operations is actually just “censorship” with various forms of cover, whether academic or operational.

Then Elon Musk rolled out a feature often used by trust & safety teams internally, looking at where accounts were created and/or where they normally post from. Except Musk made the info public. And within hours it revealed exactly why platforms had been doing this kind of work internally in the first place.

And, no, it wasn’t about “censorship” of ideological viewpoints.

On Friday, X began rolling out a feature that revealed where users signed up from and where they were posting from. The feature came following the request of MAGA influencer Laura Loomer, who asked Musk back in September to add country-of-origin labels to help identify foreign influence operations. We noted the irony at the time—Loomer and her friends spent years attacking the trust & safety teams who were actually working on this problem.

Whether it was because of Loomer’s request or it was already in the pipeline, Musk rolled it out.

And, within hours, the feature revealed that a ton of super popular “MAGA” accounts were actually posting from all over the globe, with large numbers posting from Eastern Europe, West Africa, or Southeast Asia.

Almost immediately after the feature launched, people started noticing that many rage-bait accounts focused on US politics appeared to be based outside of the US. Profiles with names like ULTRAMAGA🇺🇸TRUMP🇺🇸2028 were revealed to be based in Nigeria. A verified account posing as border czar Tom Homan was traced to Eastern Europe. And America_First0? Apparently from Bangladesh. An entire network of “Trump-supporting independent women” claiming to be from America was really located in Thailand.

The scale of the problem isn’t just about foreign actors exploiting the platform—it’s about Musk creating the economic incentives that make it profitable. As the popular Derek Guy account noted, Musk instituted an “engagement-based” payment structure that pays out money based on how many views, retweets, and comments you get. For people in lower income regions, trolling on politically sensitive topics in America to generate likes and clicks (especially now that they can use AI to do so) isn’t just easy—it’s an actual business model that Musk built into the platform.

This isn’t a bug in the system. It’s a feature Musk designed, then acted surprised when it attracted exactly the kind of behavior that trust & safety teams used to work to identify and limit.

While it’s fascinating that X actually rolled this out, it should raise questions about those who have spent the past few years stewing in that cesspool of foreign rage-bait manipulation, insisting it was somehow an accurate portrayal of public opinion.

Now, there are some limitations to the system. X’s head of product noted that uses of VPNs or Starlink may have the wrong country show up on their account. But VPN usage doesn’t explain accounts with thousands of posts over months, all originating from the same foreign location, all focused on American political rage-bait, all monetizing engagement through Musk’s payment system.

This is exactly the kind of coordinated inauthentic behavior that trust & safety teams were built to identify and address. Not to “censor” ideological viewpoints, but to surface when what looks like organic American political discourse is actually foreign actors gaming the system for profit.

For years, Matt Taibbi, Michael Shellenberger, and their allies have insisted that anyone working on these problems was part of a “censorship industrial complex” designed to silence political speech. Politicians like Ted Cruz and Jim Jordan repeated these lies. They treated trust & safety work as a threat to democracy itself.

Then Musk rolled out one basic feature, and within hours proved exactly why trust & safety work existed in the first place.

Will Taibbi or Shellenberger acknowledge they spent years attacking the people who were actually trying to protect the integrity of online discourse? Will they admit they helped dismantle the very systems that could have prevented what Musk just exposed?

I doubt it. They’re too busy on X, swimming in the very manipulation they insisted didn’t exist.

Filed Under: content moderation, elon musk, foreign influence, foreign influence operations, trust & safety
Companies: twitter, x

How MAGA Killed Foreign Influence Research, But Now Demands Social Media Stop Foreign Influence Campaigns

Content Moderation

from the turns-out-we-need-trust-&-safety-after-all dept

It’s fascinating how quickly the tune changes when the shoe’s on the other foot. For years, we’ve been treated to endless screaming about how any effort to identify and counter foreign manipulation on social media was “censorship” and a violation of Americans’ free speech rights. The same crowd that turned researchers into pariahs and shut down entire government offices dedicated to studying foreign influence operations are now… demanding that social media platforms identify foreign accounts?

Laura Loomer, ever the consistent voice of reason, recently posted on X:

After this week, I would like to see X put country tags on X accounts so we know which country someone is tweeting from.

Too many foreigners on X pretending to be Americans for the sake of creating discord and political violence in the US.

It’s totally out of control.

This is the same Laura Loomer who has spent years railing against any kind of trust & safety interventions as “censorship.” But now that foreign accounts might be saying things she doesn’t like, suddenly identifying foreign influence is a great idea?

Loomer’s demand perfectly encapsulates the broader MAGA approach to foreign influence operations: destroy the systems that actually work, then demand magic solutions when the problem inevitably resurfaces. And, of course, this is leaving out the basic things that anyone with an ounce of knowledge about countering foreign misinformation could tell you including (1) foreign misinfo peddlers tend not to appear in a way that lets platforms know where they’re posting from, and (2) even if they did, they’d immediately switch to VPNs or other methods of cloaking their location soon after.

Because for years, many MAGA voices have been screaming that any attempt to study or counter foreign manipulation was an assault on free speech. Remember the hysteria over the Global Engagement Center? The attacks on researchers studying disinformation? The endless, mostly baseless complaints about how Big Tech was supposedly “censoring” Americans at the behest of government agencies?

With MAGA folks in control of all aspects of the government, it was no surprise that they quickly shuttered the Global Engagement Center (GEC) with a bunch of nonsense about how it was against free speech. This ignored the reality, which was that it was created and focused on helping to suss out foreign influence campaigns and to expose them to limit their power.

You know, just like Laura Loomer is now asking to have happen.

Somewhat incredibly, just days after Loomer demanded efforts to identify foreign interference on social media, Marco Rubio’s State Department gleefully announced that the US had withdrawn from an international effort designed to help expose foreign influence campaigns that had been used for years to help expose things like secret ISIS social media campaigns.

Even as the MAGA world is now clamoring for efforts to help expose foreign influence campaigns, they’re still framing efforts that were going on before as… censorship. Here was Marco Rubio lying about the GEC:

Under the previous administration, this office, which cost taxpayers more than $50 million per year, spent millions of dollars to actively silence and censor the voices of Americans they were supposed to be serving.  This is antithetical to the very principles we should be upholding and inconceivable it was taking place in America.  

That ends today.  Under the administration of President Trump, we will always work to protect the rights of the American people, and this is an important step in continuing to fulfill that commitment.

This is simply, fundamentally untrue. GEC did not work on silencing or censoring Americans. It worked on helping to expose and counter foreign influence campaigns.

As did many trust & safety and integrity teams across the industry, many of which have been dismantled by the likes of Mark Zuckerberg and Elon Musk, despite their efforts to expose and limit malicious foreign actors. But because of a bad faith campaign (led by MAGA and MAGA-adjacent people) to paint all trust & safety and efforts to counter disinformation as “censorship” now we’re left in a space where the people who were good at exposing these campaigns are gone… but you have people like Loomer demanding that Elon Musk magically do what he’s spent the last couple of years making sure X could no longer do.

This isn’t just run-of-the-mill political hypocrisy. This represents a fundamental misunderstanding of what foreign influence operations actually look like and how they work.

So now we have the same political coalition that cheered the dismantling of foreign influence research suddenly very concerned about… foreign influence. It’s almost like they never actually cared about the principle of free speech, but rather about ensuring that the speech they liked wasn’t subject to any scrutiny.

Of course, now, the systems and infrastructure that could actually identify foreign manipulation campaigns in a sophisticated way require exactly the kind of research infrastructure that people like Jim Jordan’s House Judiciary Committee spent months attacking. The very people that Jordan and friends called the “censorship cartel” (something they never were) are exactly the folks that actually have expertise in finding foreign disinfo campaigns.

We’re left with what we’ve got: Loomer demanding that Elon Musk implement some kind of country-of-origin tagging system, apparently unaware that such systems are incredibly easy to circumvent with VPNs, and that building effective counter-measures requires exactly the kind of sustained research and institutional knowledge that her political allies just spent years dismantling.

You can’t have it both ways. You can’t spend years screaming that any effort to study foreign manipulation is “censorship” and then turn around and demand that platforms magically solve the foreign manipulation problem. Well, you can, but it makes you look like either a hypocrite or someone who never understood the issue in the first place.

The truth is, dealing with foreign influence operations is genuinely difficult and requires nuanced approaches that balance free speech concerns with the need to maintain the integrity of public discourse. It requires sustained research, institutional knowledge, and yes, sometimes it means that platforms need to make difficult decisions about what content to host and how to label it.

But instead of having that conversation, we got years of performative outrage about “censorship” that effectively neutered our ability to deal with foreign manipulation. And now, when the consequences of that approach become apparent, the same people who created this mess are demanding quick fixes.

Of course, now we’ll probably have to sit through years of these same people “discovering” every other aspect of content moderation and foreign influence operations as if they’re the first ones to think of it.

Filed Under: content moderation, countering foreign influence, disinformation, elon musk, foreign influence, foreign influence campaigns, free speech, laura loomer, marco rubio, misinformation, research, state department
Companies: twitter, x