DigitalMANIPULATIVE

Synthetic Profile Photos

What it is

Giving fake accounts AI-generated faces of people who do not exist, so that a reverse image search finds no stolen original and the persona appears to be a unique, real individual.

How it works

Fake accounts used to borrow photos of real people, and a reverse image search often exposed them. Generative adversarial networks, made widely accessible in 2019, produce unlimited photorealistic faces that match no one. Graphika and DFRLab documented what they called the first large-scale deployment in December 2019, in a network Facebook removed for coordinated inauthentic behavior. The lever is that a face triggers person perception automatically: we extend trust and individuality to a headshot. Nightingale and Farid found in 2022 that people distinguished synthetic from real faces at roughly chance and rated the synthetic ones slightly more trustworthy. Early GAN faces had tells (eyes at a fixed position in the frame, mismatched earrings, warped backgrounds), but newer image models have fewer, so artifact-hunting has a short shelf life. Meta has said the photos have not impaired its detection of networks, which rests on account behavior.

Real-world examples

  • •On December 20, 2019 Facebook removed more than 900 accounts, pages, and groups tied to The BL, which it linked to the Epoch Media Group. Graphika and DFRLab's report on the takedown found dozens of accounts with GAN-generated faces used to promote pro-Trump content.
  • •Meta's Q1 2024 Adversarial Threat Report states that threat actors, across operations originating in several countries, continue to use GAN profile photos and that this has not affected its ability to detect the networks behind them.
  • •A 2022 Stanford Internet Observatory investigation by Renée DiResta and Josh Goldstein found more than a thousand LinkedIn profiles with synthetic faces being used for commercial sales outreach, showing the tactic is not confined to politics.
  • •In 2019 the Associated Press reported on a LinkedIn persona named Katie Jones, with an apparently GAN-generated face, that had connected with Washington policy figures; experts described it as consistent with espionage tradecraft.

Ethical guidelines

  • ●A generated face presented as a real person's photo is a false identity claim. Clearly labeled avatars and illustrations are not.
  • ●Commercial use for lead generation is still deception of the person contacted, even when no politics is involved.
  • ●Do not publicly accuse someone of being fake on the basis of an odd-looking photo alone; real photos have artifacts too.

How to defend against it

  • ►Treat a clean reverse-image result as neutral, not reassuring. No matches used to suggest an original photo; now it is equally what a generated face produces.
  • ►Look beyond the headshot: does the person appear in any other photo, with other people, in a recognizable place, at different ages? Synthetic personas usually have exactly one image.
  • ►Verify through a second channel before trusting a new professional contact: check the employer's own staff page or call the organization's main number.
  • ►Artifact checks (asymmetric glasses or earrings, melted backgrounds, eyes locked at frame center) can still catch older fakes, but their absence proves nothing.
  • ►Weigh behavior over appearance: account age, connection patterns, and whether the account's activity is single-purpose.

References

  1. Graphika & Atlantic Council Digital Forensic Research Lab (2019). #OperationFFS: Fake Face Swarm. Graphika report, December 20, 2019 · link
    The first documented large-scale use of GAN-generated profile photos, in the network tied to The BL.
  2. Nightingale, S. J., & Farid, H. (2022). AI-synthesized faces are indistinguishable from real faces and more trustworthy. Proceedings of the National Academy of Sciences, 119(8), e2120481119 · link
    Near-chance human accuracy in telling synthetic from real faces and higher trust ratings for synthetic faces.
  3. Meta (2024). Adversarial Threat Report, First Quarter 2024. Meta Transparency Center, May 2024 · link
    Continued use of GAN profile photos by threat actors and the statement that this has not impaired network detection.
  4. Goldstein, J. A., & DiResta, R. (2022). Research note: This salesperson does not exist: How tactics from political influence operations on social media are deployed for commercial lead generation. Harvard Kennedy School Misinformation Review · link
    Synthetic-face LinkedIn profiles used for commercial lead generation.
Last reviewed
Suggest a correction

Detect Synthetic Profile Photos in any text

Paste any message, email, or article into our free Manipulation Detector to see if Synthetic Profile Photos or other techniques are being used on you.

Try the Detector FreeCreate Free Account

Related Articles

Dark Patterns in UX: How Apps Manipulate Your Behavior

Subscription traps, misleading interfaces, and engineered addiction. Understanding the persuasion techniques built into the apps you use every day.

7 min read

OSINT for Beginners: Open Source Intelligence Explained

Open Source Intelligence (OSINT) uses publicly available data to gather actionable insights. Here is a beginner-friendly guide to what OSINT is and how it is used.

10 min read