An Iraqi court gave six years for AI-fabricated blackmail images, under a penal code written before the internet

What happened

In the week before 7 October 2025, the al-Rusafa Criminal Court in Baghdad sentenced a man to six years in prison for threatening to publish fabricated images of a young woman, generated with AI. Shafaq News reported the conviction and the articles it rested on.

The conviction relied on Article 430/1 of the Iraqi Penal Code, alongside Articles 47, 48 and 49, which cover threats and complicity. Iraq has no provision addressing crimes facilitated by artificial intelligence, so these cases are interpreted under existing statutes on fraud, defamation and extortion, and Shafaq reports the judges themselves pointing at that limitation. Articles 430 and 433, on threats, defamation and extortion, are what the judiciary works with.

Furat Al-Azzawi, a legal advisor, told Shafaq: “The law was written in the mid-20th century, at a time when the internet and AI did not exist. Applying old categories to new crimes leaves loopholes.”

The offence was the threat, not the image

Nothing in Shafaq’s account of the case involves proving how the picture was made. The conduct charged was threatening to publish it, and the articles cited are the ordinary ones about threats and complicity.

That is our reading rather than something the report spells out, and it points somewhere practical. The instinct when someone is targeted this way is to prove the image is fake, which is slow, technical, and often impossible with the tools an Iraqi NGO has. The record that the law already recognises is much easier to collect: the messages, the demand, the account they came from, the timestamps.

There is no product to blame

Shafaq names no app. The tools are “free or low-cost smartphone apps”. The article also describes voice cloning built from short audio samples, and open-source collection of old photographs and family details, used to fabricate conversations and screenshots that look genuine.

So there is nothing here to block, ban or detect at the source. Anyone’s phone is enough. That removes a comfortable answer, and it is worth saying out loud before someone proposes a technology solution at a workshop.

The second harm

Shafaq is direct about why this lands harder in Iraq. Where family honour is tied to women’s behaviour, activists warn that victims may face honour-based violence or social isolation, which often leaves them little choice but to pay. The Iraqi Women Network’s 2024 report records anxiety, depression and suicidal thoughts among victims of digital blackmail, and people abandoning their online presence entirely to protect themselves.

The woman quoted anonymously by Shafaq put it plainly: even if the pictures were false, people would believe them, and she could not see how to explain otherwise.

Read that as a publication rule. A woman being extorted with a fabricated image is at risk from the reaction of the people around her, not only from the person extorting her. Case studies, awareness posts and funding reports that add identifying detail, or describe the image, move the risk toward her.

What to do differently

Collect the threat, not the image. Screenshots of the messages and the demand, the account, the dates, stored somewhere she controls and can hand to a lawyer. That is the material the existing law already recognises, and it does not require anyone to prove anything about how the file was generated.

Civil society groups quoted by Shafaq are asking for a cybercrime law that covers AI-driven abuse and online harassment of women. The gap is real. It is not a reason to wait.