---
type: source
id: bray-2023-deepfake-faces
title: Testing human ability to detect 'deepfake' images of human faces
authors:
- Sergi D. Bray
- Shane D. Johnson
- Bennett Kleinberg
year: 2023
venue: Journal of Cybersecurity
volume: 9(1)
pages: tyad011
doi: 10.1093/cybsec/tyad011
peer_review: peer-reviewed-journal
evidence_tier: B
design: 'Online experiment (Prolific); 20 images per participant drawn from 50 fake + 50 real; four conditions: control, familiarisation (20 example deepfakes), one-time advice (10 tell-tale features), advice + reminders'
modality:
- image-face
stimuli: Uncurated StyleGAN2 (FFHQ-trained) faces vs real FFHQ faces
n_participants: 280
n_note: ''
topics:
- human-detection
- training
- overconfidence
summary: Accuracy was 62.18%, barely above chance; advice and familiarisation did not significantly help; confidence was high (6.83/10) and unrelated to accuracy.
results:
- metric: Overall accuracy
  value: 62.18%
  note: '''only just above chance'''
  source_location: null
- metric: By condition
  value: control 59.65%, familiarisation 60.54%, one-time advice 64.15%, advice+reminders 64.26%
  note: differences n.s. (ANOVA p>0.1)
  source_location: null
- metric: Per-image accuracy
  value: ≈30% to ≈85%
  note: ≈1 in 5 images correct <50% of the time
  source_location: null
- metric: Mean confidence
  value: 6.83 / 10
  note: within-person gamma ≈0.06-0.21; between-person n.s.
  source_location: null
moderators:
- Interventions raised accuracy on fakes but lowered it on real images (more real faces called fake) — a criterion shift, not better discrimination.
limitations:
- 20 images per participant
- Online sample
- StyleGAN2 only
verification:
  status: verified-fulltext
  confidence: high
  checked: 2026-10-07
  via: academic.oup.com article page
url: null
pmid: null
pmcid: null
arxiv_id: null
openalex_id: null
oa_url: null
cited_in:
- MR-0001
updated: '2026-10-07'
---

# Testing human ability to detect 'deepfake' images of human faces

> **Source record** `bray-2023-deepfake-faces` · 2023 · Journal of Cybersecurity  
> **Evidence tier:** B (peer-reviewed-journal) · **Verification:** verified-fulltext  
> **Bottom line:** Accuracy was 62.18%, barely above chance; advice and familiarisation did not significantly help; confidence was high (6.83/10) and unrelated to accuracy.

## Citation {#citation}

Sergi D. Bray, Shane D. Johnson, Bennett Kleinberg (2023). Testing human ability to detect 'deepfake' images of human faces. *Journal of Cybersecurity*, 9(1), tyad011. https://doi.org/10.1093/cybsec/tyad011

Verification entry (identifiers, resolvers, claims with source locations): [/data/citations.json#bray-2023-deepfake-faces](/data/citations.json#bray-2023-deepfake-faces)

## Study design {#design}

| Field | Value |
|---|---|
| Design | Online experiment (Prolific); 20 images per participant drawn from 50 fake + 50 real; four conditions: control, familiarisation (20 example deepfakes), one-time advice (10 tell-tale features), advice + reminders |
| Modality | image-face |
| Stimuli | Uncurated StyleGAN2 (FFHQ-trained) faces vs real FFHQ faces |
| Participants (N) | 280 |
| Topics | human-detection, training, overconfidence |

## Results {#results}

| Metric | Value | Note |
|---|---|---|
| Overall accuracy | 62.18% | 'only just above chance' |
| By condition | control 59.65%, familiarisation 60.54%, one-time advice 64.15%, advice+reminders 64.26% | differences n.s. (ANOVA p>0.1) |
| Per-image accuracy | ≈30% to ≈85% | ≈1 in 5 images correct <50% of the time |
| Mean confidence | 6.83 / 10 | within-person gamma ≈0.06-0.21; between-person n.s. |

## Moderators {#moderators}

- Interventions raised accuracy on fakes but lowered it on real images (more real faces called fake) — a criterion shift, not better discrimination.

## Limitations {#limitations}

- 20 images per participant
- Online sample
- StyleGAN2 only

## Cited in {#cited-in}

- [MR-0001](/metareports/human-detection-accuracy.md): Human Detection of Deepfakes: Accuracy, Bias, Confidence, and Training

## Verification {#verification}

| Field | Value |
|---|---|
| Status | `verified-fulltext` |
| Compiler confidence | high |
| Checked on | 2026-10-07 |
| Via | academic.oup.com article page |