Can scammers really fake a video call using AI?

Yes. Scammers use face-swap deepfake technology to conduct convincing video calls. Signs include lip-sync issues, jerky movements, and a face that blurs at the edges when moving quickly. They often claim bad connection to hide quality problems.

How can I tell if a voice message was generated by AI?

Listen for robotic intonation, lack of breathing sounds, unusual pauses, and flat emotional delivery. Ask the person to say a specific phrase that includes a number — AI voice clones often mispronounce digits or sound unnatural.

What is the best way to test if I am talking to a real person or a chatbot?

Ask about something specific you discussed days ago. AI chatbots have no memory of prior conversations and will give generic or evasive answers. You can also copy their messages into GPTZero or Copyleaks to detect AI-written text.

Are AI-generated profile photos easy to detect with a reverse image search?

No — AI-generated faces are unique and will not appear in reverse image search results. You need dedicated AI-detection tools such as Hugging Face's AI or Not, Sensity, or a professional AI Image and Deepfake Review service.

AI Fraud Alert

The Role of AI in Modern Russian and Ukrainian Romance Scams

By Michael Muinov, Founder and Lead Investigator · Active casework since 1999 · Updated 17 May 2026

Artificial intelligence has changed the game. Scammers now use AI‑generated faces, voice cloning, and chatbot scripts to create entirely synthetic people. Here’s what you need to know.

Quick answer

How is AI changing Russian and Ukrainian romance scams in 2026?

Three concrete shifts have happened in the last 24 months. AI-generated faces replace stolen photos: scam operations now use fully synthetic profile pictures from StyleGAN, This Person Does Not Exist, and similar generators — meaning reverse image search returns no matches, which used to be the strongest single warning sign. LLM-driven conversation: ChatGPT-class models produce smooth, emotionally responsive English in real time, eliminating the broken-English signal that exposed thousands of scams in the 2010s. Deepfake video on demand: live video calls can now be staged with a real operator behind a real-time face-swap or generated avatar, defeating the once-decisive "ask her to video call" test.

Important limit: AI has raised the floor for what convincing fraud looks like, but it has not eliminated the fundamentals. AI cannot create a registered identity in Russian or Ukrainian state records. It cannot produce a verifiable employment history, an existing residence, or a real family. Verification has shifted from "does she look real?" to "does she exist on paper?" — and the second question is now the only one that matters.

AI‑Generated Faces (GANs)

Websites like “This Person Does Not Exist” demonstrate how realistic synthetic faces have become. Scammers use these to build profiles. Look for:

Unnatural symmetry and perfect skin texture.
Inconsistent background blur.
Strange artefacts around ears, eyes, or hair.

Our AI Image & Deepfake Review can manually detect these telltale signs.

Voice Cloning & Deepfake Video

Scammers can clone a voice from a short audio clip and even create convincing video calls using face‑swap technology. They may claim technical problems to obscure the quality. If a video call feels off – jerky movements, lip‑sync issues – trust your instincts.

Chatbots Writing Love Letters

Some scammers use AI chatbots (like ChatGPT) to generate romantic messages, making them appear eloquent and thoughtful. The problem is that the chatbot can’t answer specific, personal questions about your past conversations.

Test: Ask about something unique you discussed days ago. If the response is generic or evasive, you’re likely talking to a bot or a scammer using a bot.

How to Protect Yourself

Verify every profile. Use a Catfish Investigation to uncover the real person behind the screen. Check our Scam‑Risk Review and the Russia/Ukraine hubs. See also How to Catch a Catfish in 2026.

How to Detect AI‑Generated Identities in Romance Scams

Inspect profile photos for AI artifacts. Use AI detection tools: Hugging Face’s “AI or Not”, Sensity, or Maybe’s AI Art Detector. Look for unnatural eyes (mismatched reflections), asymmetrical glasses, blurry teeth, or distorted backgrounds.
Listen for voice cloning in audio messages. If she sends voice notes, listen for robotic intonation, lack of breathing sounds, or unusual pauses. Ask her to say a specific phrase that includes a number (AI voices often mispronounce numbers).
Analyze message text with AI detectors. Copy her messages into GPTZero, Copyleaks, or Originality.ai. AI‑written texts are often perfectly grammatically correct but lack personal details, have repetitive structure, and avoid contractions.
Test with a culturally specific question that AI cannot answer. Ask: “What is the name of the banya district in your city?” or “How do you make real borscht – with or without beans?” AI models give generic answers, while a real Russian/Ukrainian knows instantly.
Demand a live, unprompted video showing room panning. Ask her to slowly pan her phone camera around the room on a live video call. AI deepfakes cannot generate consistent, real‑time 3D environments. If the call is “always broken” or the video is pre‑recorded, it is a scam.
Check for repeating phrases across conversations. Scammers using ChatGPT often reuse identical emotional phrases (“my heart is singing when I talk to you”). Keep a log – if you see the same sentence twice, it is a script.
Use metadata analysis on image files. Download a photo she sent and check EXIF data (using tools like Jeffrey’s Exif Viewer). AI‑generated images often have no metadata or show “Created by AI” in software fields. Real smartphone photos have camera make, model, and GPS (if enabled).

Don’t let AI fool you.

Get a Professional Verification