Seeing is no longer believing

Cover of Seeing is no longer believing, Frontier Report No. 07

Frontier Report No. 07 · 16 pages · 10 sources

Can people tell AI-generated faces from real ones?

No. In one study, people judging whether faces were real or AI-generated were right 48 per cent of the time, no better than chance, and rated the synthetic faces as slightly more trustworthy. Training raised accuracy only to 59 per cent, and people shown deepfake videos overestimated their ability to spot them.

What the report found

  1. People can no longer tell real faces from fake ones. In one study, people judging whether faces were real or AI-generated were right 48 per cent of the time, no better than chance, and rated the synthetic faces as slightly more trustworthy.
  2. They think they can. People shown real and fake videos could not reliably spot the fakes, leaned towards believing they were real, and overestimated their own ability.
  3. Fakes can now talk back. A cloned voice or a faked video call can ask for money in the name of someone the listener trusts. The UAE Cybersecurity Council has urged people to verify such content through official sources.
  4. People and machines do better together. Ordinary people were about as accurate as a leading detection model on some videos, and combining the two did better than either.
  5. The bigger risk may be doubt. Labelling headlines as AI-generated made people distrust them even when they were true, and fake videos can leave viewers uncertain rather than deceived, lowering trust in all news.

What is inside

  1. What is new
  2. What older research tells us
  3. What early studies suggest
  4. What nobody knows yet
  5. Questions worth watching

Plus the UAE's response, a summary table and a graded list of claims.

Who it is for

For banks, companies, platforms and anyone who has ever trusted a familiar voice on the phone.

How we work

Frontier Reports look at behaviour the textbooks have not yet caught up with. Every claim is graded Established, Emerging or Untested. Where researchers disagree, as they do on whether the main danger is deception or doubt, both sides are set out before we say where we stand.

Cite this report

The Why We Project (2026) Seeing is no longer believing: Synthetic media and persuasion. Frontier Report No. 07. Dubai: Why We Tech Ltd. Available at: thewhyweproject.com