CoChat
Sign Up
Cited by 29Misinformation and Its ImpactsDigital Media Forensic DetectionSpam and Phishing Detection

Perception and deception: Exploring individual responses to deepfakes across different modalities

Open access
Ahmed, Saifuddin, Chua, Hui Wen·2023·Heliyon
DOI 10.1016/j.heliyon.2023.e20383PMC PMC10556585
Open this paper in CoChat — free accountOpen PDF

Abstract

This study is one of the first to investigate the relationship between modalities and individuals' tendencies to believe and share different forms of deepfakes (also deep fakes). Using an online survey experiment conducted in the US, participants were randomly assigned to one of three disinformation conditions: video deepfakes, audio deepfakes, and cheap fakes to test the effect of single modality against multimodality and how it affects individuals' perceived claim accuracy and sharing intentions. In addition, the impact of cognitive ability on perceived claim accuracy and sharing intentions between conditions are also examined. The results suggest that individuals are likelier to perceive video deepfakes as more accurate than cheap fakes, but not audio deepfakes. Yet, individuals are more likely to share video deepfakes than cheap and audio deepfakes. We also found that individuals with high cognitive ability are less likely to perceive deepfakes as accurate or share them across formats. The findings emphasize that deepfakes are not monolithic, and associated modalities should be considered when studying user engagement with deepfakes.

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Integrity signals

Registry

Crossref · 10.1016/j.heliyon.2023.e20383

Cites

10 of 10 references verified0 citation errors · 0 not checkable

Statistics

Claims the paper makes

The claims and statistics of this paper haven’t been checked yet.

Run the check — free, no account needed. It reads the open-access text and checks the sources it cites.

A few free checks a day · papers already verified open instantly · results are public

Claims other papers make about this one

Each is checked against this paper’s own text: does it say what the citing paper says it does?

  • Ethnicity has no effect on deepfake detection ability.

    “Ethnicity has no e ff ect on detection ability [42]”

    From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [42]

    Table 1 in the Results shows no statistically significant association between the study's White-versus-non-White race variable and perceived claim accuracy. But the study measured belief in a deepfake's claim, not deepfake detection ability, and a nonsignificant result does not establish that ethnicity has no effect. (LLM, conf=0.85)

    Experimental — accuracy not yet measured

  • Overall, deepfake detection does not seem to vary based on gender, education, income, or profession.

    “Overall, deepfake detection does not seem to vary based on gender, education, income, or profession [7, 32, 37].”

    From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [42]

    The Results (Table 1; page unavailable) show no statistically significant associations of gender, education, or income with perceived claim accuracy. But the study measured perceived accuracy of a political deepfake’s claim, not deepfake detection directly, and its demographic controls did not include profession. The source therefore does not support the full claim. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • Higher-quality deepfakes are associated with fewer visual artifacts, making them more difficult to detect.

    “Higher-quality deepfakes are associated with fewer visual artifacts, making them more di ffi cult to detect [20, 35, 36, 42].”

    From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [42]

    The Introduction says visual mismatches such as poor lip-syncing and unnatural expressions can signal inauthenticity, and the Discussion reports that viewers perceived video deepfakes as more accurate than cheapfakes. It suggests that a well-made deepfake may lack the dissonant cues that made the cheapfake seem false. It does not establish a general relationship between quality, the number of visual artifacts, and detection difficulty across deepfakes; its comparison is primarily with cheapfakes. (LLM, conf=0.83)

    Experimental — accuracy not yet measured

  • Demographic factors such as gender, education, and income level appear to have no discernible effect on detection performance.

    “Demographic factors such as gender, education, and income level appear to have no discernible e ff ect on detection performance [7, 32, 37, 42]”

    From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [42]

    The Results table (page not available) shows no statistically significant associations for gender, education, or income with perceived claim accuracy. But the study measured ratings of a political misinformation claim, not participants’ performance at detecting whether content was a deepfake; that part of the citing sentence is not established. (LLM, conf=0.86)

    Experimental — accuracy not yet measured

  • Lower-quality deepfakes are generally easier to detect.

    “Second, while lower-quality deepfakes are generally easier to detect [20, 35, 36, 42], imperfections in authentic videos (e.g., lighting inconsistencies or unusual framing) can lead viewers to miscategorize genuine content as manipulated [2, 22, 27, 35, 37].”

    From Human Performance in Deepfake Detection: A Systematic Review (2025) · cites this paper as [42]

    The Introduction says conspicuous audio-visual mismatches in cheapfakes reduce believability, and the Discussion reports that viewers perceived video deepfakes as more accurate than cheapfakes. But the paper treats cheapfakes as a distinct form of manipulation; it does not establish that lower-quality deepfakes are generally easier to detect. (LLM, conf=0.84)

    Experimental — accuracy not yet measured

Self-check

Percentages, sums and reported differences re-computed from the paper’s own numbers.

The self-check needs the paper’s full text.

Cites

Flagged first, then by citations

Research works printed but not linked: 31

  1. [1]Kim Y. Chen H.-T. Wang Y. Living in the smartphone age: examining the conditional indirect effects of mobile phone use on political participation J. Broadcast. Electron. Media 60 4 2016 694 713 10.1080/08838151.2016.1203318 · DOI
  2. [2]Beam M.A. Hutchens M.J. Hmielowski J.D. Clicking vs. sharing: the relationship between online news behaviors and political knowledge Comput. Hum. Behav. 59 2016 215 220 10.1016/j.chb.2016.02.013 · DOI
  3. [5]Westerlund M. The emergence of deepfake technology A Review. Technology Innovation Management Review 9 11 2019 40 53 10.22215/timreview/128 · DOI
  4. [6]Hancock J.T. Bailenson J.N. The Social Impact of Deepfakes | Cyberpsychology, Behavior, and Social Networking 2021 https://www.liebertpub.com/doi/full/10.1089/cyber.2021.29208.jth 33760669 · DOI · Paper page
  5. [7]Chi H. Maduakor U. Alo R. Williams E. Integrating deepfake detection into cybersecurity curriculum Proceedings of the Future Technologies Conference 2020 Springer Cham 588 598
  6. [9]Pashentsev E. The malicious use of deepfakes against psychological security and political stability The Palgrave Handbook of Malicious Use of AI and Psychological Security 2023 Springer International Publishing Cham 47 80
  7. [10]Sareen M. Threats and challenges by DeepFake technology DeepFakes 2022 CRC Press 99 113
  8. [16]Pennycook G. Rand D.G. Lazy, not biased: susceptibility to partisan fake news is better explained by lack of reasoning than by motivated reasoning Cognition 188 2018 39 50 10.1016/j.cognition.2018.06.011 29935897 · DOI · Paper page
  9. [19]Gurban M. Thiran J.-P. Drugman T. Dutoit T. October 20). Dynamic Modality Weighting for Multi-Stream Hmms Inaudio-Visual Speech Recognition | Proceedings of the 10th International Conference on Multimodal Interfaces 2008 https://dl.acm.org/doi/abs/10.1145/1452392.1452442?casa_token=xiikZknpb84AAAAA:g66fwO_Aobg07TkOFf-XK9wy27G7nRjboJNyADLVfRnSxndHqUMH_lTEgj6oKjzl_CUKbunwCKqo · DOI
  10. [20]Mittal T. Bhattacharya U. Chandra R. Bera A. Manocha D. M3er: multiplicative multimodal emotion recognition using facial, textual, and speech cues Proc. AAAI Conf. Artif. Intell. 34 No. 02 2020, April 1359 1367
  11. [21]Bigham J.P. Jayant C. Ji H. Little G. Miller A. Miller R.C. Miller R. Tatarowicz A. White B. White S. Yeh T. VizWiz | Proceedings of the 23nd Annual ACM Symposium on User Interface Software and Technology 2010
  12. [22]Swerts M. Krahmer E. Facial expression and prosodic prominence: effects of modality and facial area J. Phonetics 36 2 2008 219 238 10.1016/j.wocn.2007.05.001 · DOI
  13. [24]Navarretta C. Individuality in communicative bodily behaviours Esposito A. Esposito A.M. Vinciarelli A. Hoffmann R. Müller V.C. Cognitive Behavioural Systems 2012 Springer 417 423 10.1007/978-3-642-34584-5_37 · DOI
  14. [25]Kamachi M. Hill H. Lander K. Vatikiotis-Bateson E. Putting the face to the voice': matching identity across modality Curr. Biol. 13 19 2003 1709 1714 14521837
  15. [27]Wahn B. König P. Is attentional resource allocation across sensory modalities task-dependent? Adv. Cognit. Psychol. 13 1 2017 83 96 10.5709/acp-0209- PMC5405449 28450975 · DOI
  16. [30]Grimes T. Mild auditory-visual dissonance in television news may exceed viewer attentional capacity Hum. Commun. Res. 18 2 1991 268 298 10.1111/j.1468-2958.1991.tb00546.x · DOI
  17. [31]Oleszkiewicz A. Karwowski M. Pisanski K. Sorokowski P. Sobrado B. Sorokowska A. Who uses emoticons? Data from 86702 Facebook users Pers. Indiv. Differ. 119 2017 289 295 10.1016/j.paid.2017.07.034 · DOI
  18. [32]Apple W. Streeter L.A. Krauss R.M. Effects of pitch and speech rate on personal attributions J. Pers. Soc. Psychol. 37 5 1979 715 727 10.1037/0022-3514.37.5.715 · DOI
  19. [33]Carlo G. Mestre M.V. McGinley M.M. Samper P. Tur A. Sandman D. The interplay of emotional instability, empathy, and coping on prosocial and aggressive behaviors Pers. Indiv. Differ. 53 5 2012 675 680 10.1016/j.paid.2012.05.022 · DOI
  20. [34]Shields S.A. The politics of emotion in everyday life Appropriate” Emotion and Claims on Identity 2005 https://journals.sagepub.com/doi/abs/10.1037/1089-2680.9.1.3 · DOI · Paper page
  21. [36]Ahmed S. Ng S.W.T. Bee A.W.T. Understanding the role of fear of missing out and deficient self-regulation in sharing of deepfakes on social media: evidence from eight countries Front. Psychol. 14 2023 1127507 PMC10027756 36959999
  22. [37]Plomin R. Genetics and general cognitive ability Nature 402 6761 1999 C25 C29 10.1038/35011520 10591222 · DOI
  23. [38]Fiske S.T. Taylor S.E. Social Cognition: from Brains to Culture 2013 SAGE
  24. [44]Gonzalez C. Thomas R.P. Vanyukov P. The relationships between cognitive ability and dynamic decision making Intelligence 33 2 2005 169 186 10.1016/j.intell.2004.10.002 · DOI
  25. [45]Ståhl T. van Prooijen J.-W. Epistemic rationality: skepticism toward unfounded beliefs requires sufficient cognitive ability and motivation to be rational Pers. Indiv. Differ. 122 2018 155 163 10.1016/j.paid.2017.10.026 · DOI · Paper page
  26. [50]Newman E.J. Garry M. Unkelbach C. Bernstein D.M. Lindsay D.S. Nash R.A. Truthiness and falsiness of trivia claims depend on judgmental contexts J. Exp. Psychol. Learn. Mem. Cognit. 41 5 2015 1337 1348 10.1037/xlm0000099 25822783 · DOI
  27. [51]Nelson D.L. Reed V.S. Walling J.R. Pictorial superiority effect J. Exp. Psychol. Hum. Learn. Mem. 2 5 1976 523 528 10.1037/0278-7393.2.5.523 1003125 · DOI
  28. [52]Sundar S.S. The MAIN Model: A Heuristic Approach to Understanding Technology Effects on Credibility 2008 MacArthur Foundation Digital Media and Learning Initiative Cambridge, MA 73 100
  29. [55]Ferrey A.E. Burleigh T.J. Fenske M.J. Stimulus-category competition, inhibition, and affective devaluation: a novel account of the uncanny valley Front. Psychol. 6 2015 PMC4358369 25821439
  30. [57]Powell T.E. Boomgaarden H.G. De Swert K. de Vreese C.H. A clearer picture: the contribution of visuals and text to framing effects J. Commun. 65 6 2015 997 1017 10.1111/jcom.12184 · DOI
  31. [59]Itti L. Baldi P. Bayesian surprise attracts human attention Vis. Res. 49 10 2009 1295 1306 10.1016/j.visres.2008.09.007 18834898 PMC2782645 · DOI · Paper page

Cited by

Sorted by citations

Authors
Ahmed, S., & Chua, H. W.
Affiliations
Nanyang Technological University (SG)
Corresponding author
Ahmed, Saifuddin
ORCID
Ahmed, Saifuddin
Venue
Heliyon
Publisher
Elsevier BV
Published
2023-10
Volume
9
Issue
10
Pages
e20383
ISSN
2405-8440 (ISSN-L)
DOI
10.1016/j.heliyon.2023.e20383
PMC
PMC10556585
OpenAlex
W4386914235
Cited by
29
References
10
Citations by year
2026: 13 · 2025: 11 · 2024: 5
FWCI
12.46
Type
journal article
Language
en
Open access
Licence
Locations
  • Heliyon (publishedVersion, cc-by-nc-nd)
  • PubMed Central (publishedVersion, cc-by-nc-nd)
  • PubMed (publishedVersion)
  • DOAJ (DOAJ: Directory of Open Access Journals) (submittedVersion, cc-by-sa)
  • DR-NTU (Nanyang Technological University) (submittedVersion, cc-by-nc-nd)
Funders
Nanyang Technological University
Topics
  • Social Sciences › Sociology and Political Science › Misinformation and Its Impacts
  • Physical Sciences › Computer Science › Computer Vision and Pattern Recognition › Digital Media Forensic Detection
  • Physical Sciences › Computer Science › Information Systems › Spam and Phishing Detection
Keywords
deepfakes, disinformation, multimodality, deepfake audio, belief in misinformation, sharing intention, cognitive ability, deception

About these checks

Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.

Statistics and arithmetic are recomputed, not judged.

Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.

What counts as a claim ›How we measure confidence ›How we check this ›

Are you an author, or do you see a problem with this page? Request a review