A new federal case shows how chats with AI systems can become key evidence in criminal investigations. Court records reveal that the Federal Bureau of Investigation (FBI) obtained a search warrant ordering X to provide records tied to prompts used with its AI chatbot. Investigators say those prompts helped generate hundreds of nonconsensual sexual videos that targeted a real woman.
The affidavit centers on Simon Tuck, who faces accusations of harassment, cyberstalking, and intimidation against a woman he knew and her husband. According to investigators, Tuck had regular contact with the woman.
He worked out with her and exchanged text messages. During that time, agents allege he secretly recorded video of her while she exercised in his garage.
Authorities say the harassment campaign grew over several months and moved far beyond online abuse. The affidavit claims Tuck swatted the couple’s home, triggering a police response based on false emergency reports.
He also sent anonymous complaints to the husband’s employer, accusing him of child abuse and drug use.
FBI Uncovers Campaign of Impersonation and Deepfake Creation
Investigators say he impersonated the husband to send threats involving mass violence and suicide. In another incident, he contacted a funeral home and warned that the husband would soon be dead. Messages sent to the victim also claimed to come from a supposed Russian hacking group called Sector 16.
Federal agents say AI tools played a central role in the case. In January, investigators secured a warrant for conversations linked to the suspect’s use of the chatbot.
The FBI states that it received prompts that produced roughly 200 pornographic videos depicting a woman who closely resembled the victim’s wife.
One example included detailed instructions describing a staged scenario in which a blonde athlete undressed on a tennis court. The prompt specified clothing, body type, and actions, ending in explicit nudity. Investigators argue that the repeated prompts show intent to create sexual material without the woman’s consent.
The affidavit also alleges that the suspect used the chatbot to draft a formal complaint about the husband, which was then submitted to his workplace. Agents say this shows how generative AI tools can assist both harassment and impersonation efforts when used with malicious intent.
AI Logs as Evidence in Modern Harassment Cases
While the alleged conduct stands out for its scale and severity, experts note that the pattern itself is familiar. Victims of stalking and harassment often face coordinated campaigns that mix online abuse with real-world intimidation.
What makes this case different is the clear role of AI chat logs as evidence. Law enforcement now treats conversations with chatbots much like emails, text messages, or social media posts.
The case also raises questions about platform responsibility. Investigators say the videos were generated during a period when the chatbot faced criticism over weak content safeguards. Online users had already reported cases in which the system produced explicit or abusive material when prompted in certain ways.
Critics argue that gaps in moderation allowed users to push the tool into creating harmful content that targeted identifiable people.
Legal experts say the investigation signals a shift. AI companies may now receive more warrants seeking user prompts and generated outputs. Courts have long allowed access to digital communications when tied to alleged crimes. AI conversations appear to fall into the same category.
For victims, the stakes are personal and immediate. Nonconsensual sexual imagery can spread fast, damage reputations, and cause lasting emotional harm. When combined with threats and impersonation, it can create fear that extends beyond the internet into daily life.
The affidavit does not decide guilt, and the allegations remain subject to court proceedings. Still, the case offers a clear warning. As generative AI tools become common, the records they create may also become powerful evidence. What users type into an AI system can leave a trail, and investigators are now willing to follow it.
Comments are closed.