A 36-year-old man from Commerce Township, Michigan, is facing federal charges for allegedly using artificial intelligence to create explicit images of a social media influencer. The man, identified as Josh Stilman, is accused of cyberstalking the woman and threatening to publish the AI-generated images online if she did not comply with his demands.
According to court documents released on September 16, 2025, Stilman has been charged with interstate communications of a threat to extort and cyberstalking. The case highlights a growing concern over the malicious use of AI technology to harass and intimidate individuals online.
Key Takeaways
- Josh Stilman, 36, of Oakland County, faces federal charges for cyberstalking and extortion threats.
- He is accused of creating and sending AI-generated nude images of a social media influencer.
- The investigation began after the victim reported threatening Instagram messages to the FBI in March 2025.
- Evidence was gathered through search warrants for Stilman's Instagram and Google accounts.
- The case underscores the significant personal and safety impacts of digital harassment on victims.
Federal Charges Filed in Oakland County
Josh Stilman was formally charged after an extensive investigation by the Federal Bureau of Investigation. The charges, interstate communications of a threat to extort and cyberstalking, are federal offenses due to the use of online platforms to communicate threats across state lines.
The investigation was initiated in March 2025 when the FBI's Portland office received a complaint from a social media influencer. She reported receiving unsolicited and threatening direct messages on Instagram from an account with the username “FriendBlender.”
Court records show that Stilman was released on a $10,000 bond following his arrest. He is scheduled to appear in federal court again on October 14, 2025.
The Rise of AI-Generated Content in Harassment
This case is part of an alarming trend involving the use of AI tools to create non-consensual explicit images, often referred to as "deepfake pornography." These tools can manipulate existing photos to produce highly realistic but entirely fabricated images, which are then used for blackmail, harassment, or public humiliation. Law enforcement agencies are increasingly encountering cases where this technology is weaponized for criminal purposes.
Details of the Alleged Harassment
The victim, a social media influencer with nearly 100,000 followers at the time, initially received messages containing AI-generated nude images of herself and other women. According to the court filing, she generally does not respond to unsolicited messages but replied on one occasion in an attempt to dismiss the sender.
The messages reportedly stopped for a period but resumed several months later with a more aggressive and threatening tone. The “FriendBlender” account allegedly demanded that she respond and threatened to post the pornographic images publicly if she refused.
Escalating Threats via Direct Message
The court documents included excerpts of the messages sent to the influencer, which demonstrated the nature of the alleged threats.
“I wish you’d respond to some of my questions. I really don’t want to resort to posting things online. I have much more than you’ve seen. I’m really not trying to offend you either. I j*** off to you every day, I just want to know what you like.”
The sender also sent a Google Album link which, according to the victim, contained more AI-generated explicit photos of her. This link proved to be a critical piece of evidence in the investigation.
How Investigators Identified the Suspect
The victim played a key role in identifying the suspect. She noticed that the display name associated with the Google Album link was “Josh Stilman.” This prompted her to search her own Instagram followers, where she found an account for a Josh Stilman who used the handle “ThisDIYguy.”
She recalled that this same account had messaged her in 2021. After connecting the name to the threatening messages, she confronted him directly in March 2025, asking, “Does your wife know you are messaging me?”
Immediately after her message, both the “FriendBlender” and “ThisDIYguy” accounts blocked her. The FBI reported that Stilman also deleted the Google Album and removed his LinkedIn profile around the same time. Fortunately, the victim had taken screenshots and video recordings of the messages before she was blocked, which she provided to law enforcement.
Digital Trail Leads to Arrest
Investigators used search warrants to confirm Stilman's ownership of the Instagram and Google accounts. An examination of his Google Drive history revealed searches and visits to websites specifically used for creating AI-generated pornography, directly linking him to the tools needed to create the images in question.
FBI Investigation and Confirmation
Using the information provided by the victim, the FBI in Portland began its official investigation. They identified a possible match for Josh Stilman with several addresses in the Metro Detroit area. A comparison of Stilman’s driver’s license photo with his Instagram profile picture appeared to confirm his identity.
Local police records noted previous contacts, including a mental health check call in Livonia in April 2024. FBI surveillance confirmed Stilman was residing at a home in Commerce Township.
Warrants for Stilman’s Instagram accounts confirmed he was the subscriber for both “FriendBlender” and “ThisDIYguy.” The search uncovered numerous vulgar messages and AI-generated images sent to the victim, starting on March 4, 2025.
Impact on the Victim and Potential Pattern
The alleged threats had a significant impact on the victim's life. According to her statement to authorities, she was forced to make several lifestyle changes out of fear for her safety. These changes included:
- Removing her personal address from her public website.
- Worrying constantly about the safety of her family.
- Beginning to carry a weapon for personal protection.
Furthermore, the victim told investigators she had contacted other women on social media who claimed to have experienced similar threats from Stilman, with some incidents dating back to 2019 or 2020. This suggests a potential pattern of behavior that may extend beyond this single case.