Human detection of political speech deepfakes across transcripts, audio, and video
Abstract
Recent advances in technology for hyper-realistic visual and audio effects provoke the concern that deepfake videos of political speeches will soon be indistinguishable from authentic video. We conduct 5 pre-registered randomized experiments with N = 2215 participants to evaluate how accurately humans distinguish real political speeches from fabrications across base rates of misinformation, audio sources, question framings with and without priming, and media modalities. We do not find base rates of misinformation have statistically significant effects on discernment. We find deepfakes with audio produced by the state-of-the-art text-to-speech algorithms are harder to discern than the same deepfakes with voice actor audio. Moreover across all experiments and question framings, we find audio and visual information enables more accurate discernment than text alone: human discernment relies more on how something is said, the audio-visual cues, than what is said, the speech content.
Study specs
Five pre-registered randomized experiments with varied base rates of misinformation, audio sources, question framings, and media modalities were conducted.
- Authors
- M Groh,A Sankaranarayanan,N Singh,DY Kim
- Discipline
- Computational Social Science
- Sample Size
- N=2,215
- Study Type
- Experimental Study
- Year
- 2025
- Human Data Platform
- Prolific
- Source
- View Source DOI Google Scholar
Measured Outcomes
Human accuracy in discerning real political speeches from deepfakes across media formats and contextual variables.
Peer Review & Critical Discussion
Potential Selection Bias in 2023 Cohort
The participant pool shows a concerning overrepresentation of users from high-income demographics. Looking at Table 3, we can see that 78% of respondents had annual incomes above $75k, which significantly limits the generalizability of these findings to broader populations.
Non-naive Participants Issue
I've noticed a methodological concern regarding participant naivety. Given that Prolific users often complete multiple studies, there's a real risk that participants had prior exposure to similar experimental paradigms, which could confound the results.
RLHF Applicability to This Study Design
The implications for RLHF training pipelines are understated. If we accept the authors' conclusions about preference stability, this has direct consequences for how we should structure reward model training. The temporal decay effect described in Section 4.2 is particularly relevant.
Verify your expertise to join discussion
Create an account and verify your credentials to participate in peer discussions.