Crowds Improve Human Detection of AI-Synthesised Faces

Source record kramer-2024-crowds-synthetic-faces · 2024 · Applied Cognitive Psychology
Evidence tier: B (peer-reviewed-journal) · Verification: verified-abstract (compiler confidence: medium)
Bottom line: Aggregating different people's judgements (outer crowds) improved synthetic-face detection as crowd size grew, with gains comparable to training; averaging one person's two attempts (inner crowd) added little.

Citation

Robin S. S. Kramer, Charlotte Cartledge (2024). Crowds Improve Human Detection of AI-Synthesised Faces. Applied Cognitive Psychology, 38(5). https://doi.org/10.1002/acp.4245

Verification entry (identifiers, resolvers, claims with source locations): /data/citations.json#kramer-2024-crowds-synthetic-faces

Study design

Field	Value
Design	Exp1: real-vs-synthetic ratings on a 1-7 scale, completed twice (inner and outer crowds). Exp2: single binary judgement with confidence and estimate of others' agreement (majority vote, confidence-weighted and 'surprisingly popular' aggregation)
Modality	image-face
Stimuli	AI-synthesised and real faces (details not in abstract)
Participants (N)	not recorded
N detail	not in abstract
Topics	human-detection, crowds, aggregation

Results

Metric	Value	Note
Inner crowd	little benefit over individuals
Outer crowd	performance improved with crowd size
Aggregation rule	majority vote best for small crowds; 'surprisingly popular' best for larger crowds
Personality	no association with performance

Moderators

Crowd size
Aggregation rule

Limitations

Values not checked against full text

Cited in

MR-0001: Human Detection of Deepfakes: Accuracy, Bias, Confidence, and Training

Verification

Field	Value
Status	`verified-abstract`
Compiler confidence	medium
Checked on	2026-10-12
Via	OpenAlex abstract record for doi:10.1002/acp.4245

Values on this page have not all been checked against the full text. Confirm them against the primary source before high-stakes use.