Last week, civil rights attorney Nekima Levy Armstrong was arrested after participating in a protest at a church in St. Paul, Minnesota, where the pastor had reportedly been working with ICE. The White House shared an image of Levy Armstrong following the arrest that appeared to show her crying. But the image is fake, apparently altered with AI to make her look like she was distressed or regretful. Which raises an interesting new question: What can you do if the world’s most powerful government is arresting you on trumped-up charges and then sharing fake photos of you? Do you have any recourse at all?

Nekima Levy Armstrong and Chauntyll Allen, a St. Paul school board member, were arrested Jan. 23 for violating the FACE Act, which prohibits attempts to intimidate, threaten, or interfere with services at places of worship. Video of the arrest captured by Levy Armstrong’s husband shows agents not just recording her but assuring her that the footage wouldn’t be used on social media.

“Why are you recording?” Levy Armstrong asked in the 7-minute video. “I would ask that you not record.”

“It’s not going to be on Twitter,” the unidentified agent told her. “It’s not going to be on anything like that.”

But it was posted to Twitter, now known as X. Homeland Security Secretary Kristi Noem posted an image showing Levy Armstrong with a relatively neutral expression—confident and stoic. But the X account for the White House posted something different. That account showed Levy Armstrong crying, with tears rolling down her face. It was most likely created with AI. Her lawyer, Jordan Kushner, told the Associated Press that it was defamation.

“It is just so outrageous that the White House would make up stories about someone to try and discredit them,” Kushner said. “She was completely calm and composed and rational. There was no one crying. So this is just outrageous defamation.”

Gizmodo spoke to experts to get a better sense of what Levy Armstrong could do after such an egregious move by the White House. And the consensus seems to be that any attempt to get justice will be complicated.

Eric Goldman, a law professor at Santa Clara University School of Law, pointed out that the government has been trying to crack down on malicious uses of AI to misrepresent people, yet the White House turns around and does just that, “role modeling the worst behavior that it’s trying to prevent its citizens from engaging in.”

“It’s so shocking to see the government put out a deliberately false image without claiming that they were manipulating the image. This is what we call government propaganda,” said Goldman.

Goldman says that there are several layers to a defamation claim that Levy Armstrong would need to establish to be successful.

“She’d have to show that there was a false statement of fact. And normally we treat photos as conclusive statements of fact, that they’re truthful for what they depicted, but it wouldn’t surprise me if the government argued that it was a parody or that it was so obviously false that everyone knew it was false and therefore it was not a statement of fact,” said Goldman.

“Now, that’s just sophistry, right? If defamation law means anything, it would apply to a fictionalized photo that is presented as truthful. Like, that’s what it’s supposed to cover. And yet, the government could very well win on the very first element,” Goldman continued.

A statement of fact also has to harm someone’s reputation, and that’s another hurdle, according to Goldman. We might expect someone to cry when they’re getting arrested, which means that he says it’s hard to make the case that her reputation has been harmed. There’s also the question of whether she’s a public figure.

“There’s a First Amendment defense that limits defamation claims. And they raised the bar on claims that apply to matters of public concern and public figures. And I would argue that potentially the photo subject would qualify as a public figure and her arrest was clearly a matter of public concern,” said Goldman.

Finally, she would need to show that the government had demonstrated “actual malice” about the veracity of the statements they’re making, meaning that they knew what they were presenting was false with the intent to harm her reputation. “Now if you fictionalize a photo and present it as true, I think you might have actual malice,” Goldman explained. “However, I’m not sure how that would play out in this circumstance.”

The long and the short of it? Goldman says, “It’s not clear to me that even if she sues, she wins.”

Other legal experts that Gizmodo spoke with had roughly the same response when it came right down to it. There simply isn’t a strong enough case for defamation. The remedy for the government lying about people is for the politicians in charge to get replaced.

“We’ve assumed that if politicians are gonna publish false information, the voters are gonna punish them for it,” said Goldman. “And there might’ve been a time that was true, but that model is clearly broken down.”

It’s unclear which AI image generator was used to make the crying photo. Gizmodo tested various AI chatbots to see what kind of guardrails might be in place for this kind of thing. Google’s Gemini and OpenAI’s ChatGPT made her cry. Microsoft Co-Pilot refused, as did Anthropic’s Claude, explaining, “I can’t edit images to add manipulated emotional expressions to photos of real people. This could be used to misrepresent someone or create misleading content.”

What about xAI’s Grok? The service was down when we tried. But it’s safe to say that Grok probably will let you make people cry in an attempt to ridicule them, given everything else that Elon Musk will let you do.

It’s a unique moment in modern U.S. history. The American government has been caught lying repeatedly on matters big and small as long as it has existed. But the lies of President Donald Trump’s second term are so transparently false that it’s almost laughable.

Kristi Noem got up in front of microphones on Sunday to call Alex Pretti, the man killed by ICE agents in Minneapolis, a domestic terrorist. She said that the 37-year-old ICU nurse at the VA showed up to “perpetuate violence.” It’d be amusing if it weren’t so horrifying. The government lies with impunity, and they don’t care that we all could see a compassionate and caring man murdered in the street by masked agents of the state.

When the government goes even further than mere words, attempting to manipulate the images we see with AI fakery, it somehow feels even worse, like we’re on the precipice of a post-truth society. Unfortunately, many Trump voters don’t seem to care.

“I don’t think we’ve had enough discussion about AI deepfakes being weaponized by the government’s propaganda so they can lie against their constituents,” said Goldman. “And we may not have an adequate set of resources to punish the government for such abuses.”

“I don’t know what the remedies are. I fear that we don’t have them strong enough, but I fear even more that voters are going to reward politicians for abusive propaganda. This might just be what it means to own the libs.”