Editor’s Note: A key legislative priority of First Lady Melania Trump, the recently enacted ‘Take It Down Act’ seeks to provide prosecutors and law enforcement with new legal options to help address the rising number of AI-generated nonconsensual intimate images. In this week’s The Rabbit Hole, reporter Peter Beck examines the first federal criminal cases brought under the new law and what the U.S. government is doing to protect victims of the Internet’s newest federal crime. - Seamus
A woman in Ohio ended a romantic relationship with a guy she was seeing. And then the texts started.
At first it was texts from unknown numbers. Then it was violent voicemails and images of her nude. She recognized some of the pictures she received over text as ones she had taken. But other images not so much: they portrayed her changing her clothes and using the restroom, she had never seen before. Her mother began to receive the texts and images, too.
The woman was then sent more videos and images. This time, they depicted her having sex with various men, including her own father. But she could tell the images didn’t look right; they appeared glossy and mismatched. The body of the woman intended to be her didn’t even look like hers. The images were generated through artificial intelligence (AI).
After waiting almost two weeks for the texts to stop, the woman went to the local police in Hilliard, Ohio. She told her friends about what was happening. One of them, James Bradley Strahler II, was especially sympathetic to what she was going through and offered to leave his phone on so that she could get hold of him all the time. Strahler texted her:
“I wish they would stop with you and move on. If there was a way to have them stop with you and move it to me, I would do it in a heartbeat. My heart hurts seeing this happen to you. I only brought up interacting with them only because it could be a means to an end. I’d sacrifice my life to make sure you're okay. I'm going to continue to think of options.”
However, the sympathetic support was all a ruse. In fact, Strahler went on to be the first person to be convicted under the Take It Down Act for his messages containing AI-generated harassment sent to at least six women. Passed in 2025, the act was the first major legislative push by Congress to criminalize illicit conduct—such as harassment in Strahler’s case—made more severe by AI.
Strahler’s conviction and other recent prosecutions mark a test of how federal law enforcement and prosecutors will handle the growing number of AI-enabled crimes. The cases also offer early signals of how Congress and the courts will handle the constitutional and practical challenges to policing the greatest technological development of the twenty-first century.
Strahler was questioned two days after the woman contacted the Hilliard Police Department. Law enforcement said he admitted to sending the text messages, images, and voice mails, including one in which he threatened to rape her. He was arrested, charged with a municipal offense, and served with an order of protection, prohibiting him from contacting the woman. The texts seemingly stopped.
Then, three months later in April 2025, one of the woman’s coworkers began receiving text messages with the images. At least one of the images, according to a criminal complaint in the case, Strahler had admitted during his interrogation to previously sending. Local police reportedly warned Strahler to stop contacting the woman, her family, coworkers, or friends, but the messages didn’t stop. The woman started to receive calls from phone numbers spoofing the numbers of her bosses.
Local law enforcement allegedly discovered during a forensic examination of Strahler’s phone that she was just one of three women targeted by him. Strahler had dated the other two women, too. As with the first woman, Strahler's harassment extended to their mothers as well, finding their contact information and pictures online.
Strahler purportedly used the pictures to generate media with AI of the women engaged in incestuous relationships. He sent them the manipulated media and demanded actual nude images, threatening to send the AI-generated ones to their friends and family if they didn’t comply.
He continued to escalate. “Did ur daughter [adult victim-4] get her big natural tits from your side of the family? Do you actually have big family tits or are they just deflated? Send a picture of them to me,” one text reportedly wrote. Another series of texts included in the criminal complaint stated, “Is this really you? Send pic of ur boobs or friends, family and co-workers will get this. You have hour. Tik tok Plenty to go out. Stop it all by sending ur boobs.”
IMAGE: FTC's TakeItDown Website. https://takeitdown.ftc.gov/
Meanwhile, Strahler uploaded the manipulated images of the first woman and her mom to an anonymous adult website. The same website was taken offline by Dutch authorities in May 2026 after reports that it was implicated in gender-based violence, trafficking, and assault. Twenty-three of the AI-generated images that Strahler uploaded allegedly depicted the woman and her mom engaged in inappropriate conduct with minors.
A grand jury in the Southern District of Ohio indicted Strahler for publication of digital forgeries, two counts of cyberstalking, and production of child abuse. The first of the three charges, 47 U.S.C. § 223(h)(3)(A), was the criminal offense enacted under the Take It Down Act, designed to address AI-generated and other manipulated forms of adult media. Strahler pleaded guilty to all four on July 14, making him the first in the nation to be convicted under one of the U.S.’s most recent criminal laws.
Federal prosecutors said he created more than 700 images of “both real victims and animated persons” to post to the now-removed adult website. Another 2,400 images on his phone were reportedly flagged for depicting adult content, child abuse material, or violence.
All told, prosecutors said Strahler had installed more than 24 AI platforms on his phone and accessed more than 100 AI web-based models, a number far exceeding the frontier models hosted by the leading AI firms, such as Anthropic, OpenAI, and Google. The nature of the content created and the number of models Strahler accessed likely indicate he sought out AI models specifically used to generate adult content.
In October 2023, less than a year after OpenAI’s ChatGPT publicly launched, reports began to emerge of high school students using so-called “nudify” tools to manipulate actual, innocuous images of their classmates into explicit materials. Public pressure for Congress to act mounted following the reports and brave testimony before congressional committees by minor victims and their families, including that of Elliston Berry, who was 14 when a classmate used AI to generate explicit images of her.
Welcome The Rabbit Hole, our Sunday series where we take one docket, court document, or topic touching on the federal courts and dive deep. The Rabbit Hole is typically reserved for paid subscribers of Court Watch, reflecting the time and effort required to report the story deeply. However, with the influx of hundreds of new subscribers after we uncovered a previously dormant federal court, this issue is paywall-free so our new readers can get a sense of what this series is all about. If you want to receive next week’s The Rabbit Hole, consider upgrading to a paid subscription.
In response, Senator Ted Cruz (R-Texas) introduced the Take It Down Act, which senators of both parties quickly rallied to support. The bill had two components. The first, the “take it down” portion, required covered platforms to establish a process to notify victims and remove explicit images of them published online without their consent. It also gave the Federal Trade Commission (FTC) the ability to pursue enforcement action against providers that fail to meet their obligations, and the Department of Justice the ability to seize the websites that publish the media.
The FTC has since launched a website for people to report platforms that fail to remove deepfakes. The website also provides guidance to victims on how to report websites, ask platforms to take down nonconsensual images, and contact the FBI.
On June 12, the Department of Justice and the Department of Homeland Security announced that the act has been used since its May 2025 enactment to remove adult websites that reportedly publish nonconsensual images from the internet. Officials said the websites, which were brought to the U.S. government’s attention by Italian authorities, had been used to publish digital forgeries that were manipulated to portray images of famous women, including politicians, journalists, television presenters, anchors, athletes, entertainers, royalty, and even several countries’ first ladies.
Visitors could reportedly search for videos with tags that included “rape,” “forced,” and “degradation.” The Paris Public Prosecutor’s Office, which is conducting its own investigation into the sites, arrested a 47-year-old man in Nice, France, on June 10 who allegedly helped run the sites. The two websites targeted—CFAKE.com and SOCFAKE.com—now have notices citing the Take It Down Act and declaring that the Justice Department and other federal law enforcement agencies have seized them.
As Robert Frazer, the U.S. Attorney for the District of New Jersey, which handled the seizure, said, “For the victims whose images were distributed without their consent, the harm is not virtual — it is deeply personal and often enduring.”
The second part of the act prohibited “any person, in interstate or foreign commerce, to use an interactive computer service to knowingly publish a digital forgery of an identifiable individual.” The bill defined “digital forgery” as “any intimate visual depiction of an identifiable individual created through the use of software, machine learning, artificial intelligence, or any other computer-generated or technological means.” An “identifiable individual” was to be someone “who appears in whole or in part in an intimate visual depiction” and “whose face, likeness, or other distinguishing characteristic (including a unique birthmark or other recognizable feature) is displayed in connection with such intimate visual depiction.” The penalty was capped at a two-year prison sentence for offenses involving adults and a three-year sentence for those involving minors.
Amidst the bipartisan outrage over AI’s growing misuses, the bill moved quickly through Congress compared to other legislative attempts to overhaul online speech regulation, which raise myriad thorny issues both for the left and the right. Senator Cory Booker (D-NJ) briefly slowed down the bill’s process, raising concerns about several of the act’s provisions for using vague language and potentially overbroad criminal penalties. Several civil liberty and internet advocacy groups also voiced skepticism about the bill.
Booker, however, soon reversed course amidst strong public pressure to pass the bill, allowing the Senate to pass the Take It Down Act with unanimous consent. The House passed a companion bill in April 2025, and President Trump signed it into law on May 19, 2025. Before Congress acted, almost every state had enacted at least one law regulating deepfakes.
Since the law was enacted in April 2025, the Department of Justice has filed charges in at least two other cases in addition to Strahler’s prosecution in Southern Ohio. On May 20, federal prosecutors announced separate charges against two men in the Eastern District of New York. One of them, a 51-year-old from New Jersey, allegedly uploaded AI-generated content to an adult website, depicting famous female politicians, actresses, musicians, and others. According to the criminal complaint, the FBI connected the 51-year-old’s DMV records to his profile picture on the adult site.
Law enforcement identified at least two of the politicians the man allegedly uploaded AI-generated videos of, including one based on her nickname in the article’s title. Prosecutors said the site listed his page as being visited at least 2.1 million times. A judge allowed the man to be released on a $50,000 bond before his trial.
The third defendant charged so far under the Take It Down Act, a 20-year-old from Texas, allegedly targeted famous and nonfamous victims alike. His accounts on adult websites, which prosecutors said have been viewed more than a million and a half times, allegedly advertised the content as depicting real people the account holder knew. One album reportedly included images that were manipulated from a series of innocuous graduation pictures taken from a high school yearbook to engage in sexual content. The 20-year-old allegedly published approximately 113 albums on an adult site, with at least 50 separate female victims.
So far, the Take It Down Act has provided the Department of Justice with a useful enforcement tool for the most prolific offenders, who use AI and other deepfake technology to create tens of thousands of nonconsensual images, generating millions of views on adult websites. Prosecutors have also shown a willingness to charge it in the most persistent cases, when a defendant allegedly continues to use AI to harass his victims, despite multiple interventions by law enforcement. These are the offenses that impact the largest number of victims and appear most on law enforcement’s radar.
What’s unclear, however, is how effectively the Take It Down Act will be able to be used to prevent the more isolated and targeted creation of nonconsensual adult content. As long as there are AI models available to “nudify” subjects, there will be ex-boyfriends, classmates, colleagues, and complete strangers who use the technology for nefarious reasons.
These AI applications don’t just operate in the dark corners of the internet. They advertise on the largest adult websites and regular social media platforms. On X, where its owner has encouraged users to interact with Grok—xAI’s chatbot—on “spicy mode,” users frequently leave comments on women’s innocuous pictures asking Grok to generate inappropriate images of them. And on July 16, San Francisco City Attorney David Chiu sent cease-and-desist letters to Apple and Google, demanding that the companies remove 13 “nudify” applications from their app stores. X itself also sued a Grok user this week for misusing the AI following his arrest for inappropriately exploiting minors.
As the the Justice Department begins to create a prosecutorial apparatus to address the growing number of instances in which AI has been used to create nonconsensual adult content, there will undoubtedly be a flood of new criminal cases. Strahler’s case is but the first step by prosecutors to stem the flood that is undoubtedly coming.