Pith. sign in

REVIEW 3 major objections 4 minor 1 cited by

Signals of Provenance: Practices & Challenges of Navigating Indicators in AI-Generated Media for Sighted and Blind Individuals

T0 review · 3 major / 4 minor · reviewed 2026-08-07 · deepseek-v4-flash

Challenge this review Re-run · record.json Download PDF Read on arXiv ↗

Pith's one-line read Current self-disclosed AI-content indicators fail sighted and blind users alike: in an interactive session with 12 posts, neither group reliably used platform labels, and blind participants missed visual signals such as watermarks almost…

desk verdict A useful qualitative study of how sighted and blind users navigate AI-content indicators, but the counting table overstates the case; still worth refereeing. read the letter →

arxiv 2505.16057 v2 pith:DXLH6DXU submitted 2025-05-21 cs.HC cs.AIcs.MM

Ayae Ide , Tory Park , Jaron Mink , Tanusree Sharma This is my paper
classification cs.HCcs.AIcs.MM
keywords AI-generatedcontentprovenanceindicatorsblindandlow-visionaccessibilityscreenreaderusabilitymisinformationwarninglabelssocialmediaplatformpolicyhuman-computerinteractionmentalmodelsofAI
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

This paper argues that the self-disclosed AI-generated content indicators now being rolled out by YouTube, TikTok, and Instagram are failing the people they are meant to protect. Through one-hour semi-structured interviews and an interactive screen-sharing session with 28 participants (15 sighted, 13 blind or low-vision) navigating 12 preselected AI-generated posts, the authors found that neither group reliably engaged with platform-provided menu-aided labels; instead participants looked at content-based cues such as titles, descriptions, hashtags, and comments. Blind and low-vision participants missed visual indicators such as watermarks almost entirely because screen readers and interface hierarchies do not surface them. The paper matters because regulators in the EU, China, and elsewhere are mandating exactly these disclosures; if the labels are inaccessible or overlooked, the mandates cannot deliver their intended transparency.

What carries the argument

The analytical machinery has two parts. First, a two-type taxonomy of AI indicators: content-based indicators (AI references inside ordinary post fields such as titles, descriptions, hashtags, and comments) and menu-aided AI labels (platform- or creator-supplied disclosures such as single-line labels, hidden descriptions, and rotating labels). Second, an interactive navigation protocol in which participants shared their screens while visiting 12 curated AI-generated videos, audio clips, and images on YouTube, TikTok, and Instagram, allowing the researchers to observe which indicators were noticed and used without prompting. This combination lets the study attribute failures to specific label designs and interface hierarchies rather than to general user inattention.

What would settle it

A field study that instrumented real platform usage—for example, logging whether users click or expand AI labels in the wild and testing blind users' recall of AI disclosures via screen readers—would settle the claim: if a large, diverse sample reliably noticed and correctly interpreted menu-aided labels (say, majority recall after a single exposure), the paper's central failure claim would be undercut.

Watch

Extended reading notes

Core claim

The paper's central claim is that current self-disclosed indicators do a poor job of conveying AI provenance to either sighted or blind users, and that the reasons are design-level: inconsistent placement, hidden or rotating labels, overly technical wording, and interfaces that are not structured for screen readers. The authors categorize indicators into content-based (title, description, hashtags, comments, creator watermarks) and menu-aided (platform labels such as YouTube's 'Altered or synthetic content', TikTok's single-line AI label, and Instagram's rotating 'AI info' tag). In the interactive session, engagement with menu-aided labels was low across groups and zero for blind participants with single-line labels; sighted users relied on visual and audio cues, while blind users relied on audio and assistive tools and were largely unaware of visual indicators. The paper also identifies four mental models participants use to make sense of AI-generated media—generation-oriented, identification-oriented, sensory-modality, and risk-benefit—and argues that these models explain why content-based signals are preferred: they are familiar, top-of-page, and already part of the user's scanning behavior.

Load-bearing premise

The load-bearing premise is that the behavior of 28 self-selected, mostly highly educated participants during a remote one-hour screen-sharing session with 12 preselected posts represents how sighted and blind users generally navigate AI indicators in everyday platform use; the paper's missing codebook reference also means the coding scheme behind its mental-model analysis cannot be independently checked.

Editorial extensions

If this is right

  • If the paper is right, platform-mandated AI labels in their current forms—especially hidden and rotating labels—should not be expected to reduce deception or misinformation, because most users never register them.
  • Designing indicators for screen-reader navigation, with a proper heading structure and disclosure before playback begins, would make provenance reachable for blind users who currently miss it.
  • Standardizing label placement, timing, and wording across YouTube, TikTok, and Instagram would lower the cognitive cost of finding provenance and reduce the confusion caused by inconsistent designs.
  • Regulatory requirements such as the EU Digital Services Act and AI Act should be paired with usability and accessibility standards for disclosure labels, not just with the mandate to label.
  • Policymakers and platforms treating AI disclosure as a shared responsibility—creators disclose, platforms enforce, communities report—would match user expectations better than creator-only self-disclosure.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A testable extension would be to measure whether the proposed 'canonical AI disclosure schema' and disclosure registry actually improve blind users' identification accuracy in a controlled experiment, since the paper argues for these designs but does not evaluate them.
  • The finding that sighted users also overlooked menu-aided labels suggests that usability failures, not just accessibility failures, may explain the ineffectiveness of misinformation warning labels reported in earlier work; the two problems may share a fix.
  • The BLV preference for auditory and pre-playback disclosure implies that watermarking and visual badges will continue to exclude blind users even if provenance metadata becomes universal, so provenance standards should include an audio channel.
  • If rotating labels cause users to think they have already consumed the disclosure, platforms should treat animation as an accessibility hazard rather than a feature.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

3 major / 4 minor

Summary. This paper reports a qualitative study with 28 participants (15 sighted, 13 blind or low-vision) using semi-structured interviews and an interactive screen-sharing session with 12 pre-selected AI-generated media items from YouTube, TikTok, and Instagram. The authors identify four mental models of AI-generated content, present observational counts of how often participants engaged with content-based versus menu-aided AI indicators without prompting, and derive design recommendations across content, placement, timing, modality, and responsibility. The central descriptive claim is that menu-aided platform indicators are frequently overlooked by both groups, with blind and low-vision users facing additional accessibility barriers, and that participants instead rely on content-based indicators such as titles, comments, and hashtags.

Significance. If the central claims hold, this is a valuable and under-represented contribution: it is one of the first studies to compare sighted and blind/low-vision users' practices with self-disclosed AI-content indicators, and it draws attention to accessibility failures in provenance labeling that policy discussions often ignore. The qualitative data include participant quotes and an inter-coder reliability check (0.80 on 20% of transcripts), and the paper connects its findings to concrete design proposals (e.g., a canonical disclosure schema, API-driven standardization, disclosure registries) that go beyond generic suggestions. The main risk is that the headline quantitative pattern about menu-aided labels being overlooked is not exposure-normalized, and the missing codebook prevents full verification of the coding scheme.

major comments (3)
  1. [§5.2, Table 3] The quantitative claim that both groups 'frequently overlooked' menu-aided labels is not exposure-normalized. Table 6 shows that only 4 of the 12 media items contained an AI label (items 1, 5, 6, 12), and the rotating Instagram label was described in §5.2 as available only on the mobile app while most participants joined via laptop. The paper provides no per-item or per-participant record of whether each indicator was present, rendered, or screen-reader accessible on the exact device/URL combination used. Consequently, the zero interaction counts for BLV participants with single-line and rotating labels in Table 3 may reflect non-exposure rather than overlooking. The abstract and §7.1 rely on this pattern as a central finding, so the authors should either report exposure denominators (e.g., number of participants for whom each indicator was actually available and rendered) or explicitly reframe Table 3 as describing interactions within a specific stimulus set, supported by the qualitative quotes rather than as evidence of general failure rates.
  2. [§3.4] The manuscript states 'We provided the finalized codebook (Table??) in the Appendix for reference,' but no codebook is present; the placeholder 'Table??' is unresolved. Because the inter-coder reliability value (0.80) is offered as evidence of analytic rigor, the codebook with code definitions is necessary for readers to assess what was coded and how the themes in §4–§6 were derived. The authors should include the complete codebook in the appendix and correct the cross-reference.
  3. [§3.4, §5] The behavioral observation procedure underlying Table 3 is under-specified. It is not described how 'interacted with indicator without nudging' was operationalized, whether sessions were recorded and coded from video or audio, how interaction was distinguished from incidental screen-reader traversal, or whether the counts in Table 3 were derived from the same coding process as the interview themes. This matters because Table 3 is the primary quantitative support for the 'overlooked menu-aided indicators' claim; without a clear coding protocol, the counts are difficult to interpret or reproduce.
minor comments (4)
  1. [§2.4] The text refers to 'policy review (Table 4)' but the platform policy comparison is Table 1; the cross-reference should be corrected.
  2. [Throughout] There are numerous typos and grammatical errors that should be corrected, including 'intertional' (§2.1), 'Futhermore' (§2.3), 'vcoder' (§2.2), 'navigat' (§2.3), 'rasies' (§5.2), 'ja WS' (§5.1), and 'infulencer' (Table 7).
  3. [Table 7] The 'Difficulties for Sighted' and 'Difficulties for BLV' columns use 'gray:', 'easy:', and 'difficult:' as pseudo-labels, but 'gray' is never defined; this shorthand should be replaced with explicit difficulty ratings or removed.
  4. [§3.3] The sample demographics are reported but the paper does not include a limitations subsection; the authors should discuss the self-selected recruitment (Prolific, NFB mailing list) and the skew toward post-graduate education (43%) and AI/ML backgrounds (29%) as potential bounds on transferability.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the central claims are empirical findings from interviews and screen-sharing observations, not derivations from assumptions that already contain the conclusions.

full rationale

This is a qualitative HCI interview study. Its central claims (e.g., participants frequently overlooked menu-aided AI labels and instead relied on content-based indicators such as titles, descriptions, hashtags, and comments) are grounded in observed interactions, participant quotations, and thematic coding, not in a fitted model, a mathematical derivation, or a self-citation chain. The paper does not define its outcome variables in terms of its inputs, does not fit a parameter and then relabel it as a prediction, and does not invoke a self-authored uniqueness theorem to force a design choice. Self-citations appear (e.g., reference [62] in the definition footnote and several Mink/Sharma references in related work), but none is load-bearing: the empirical findings stand on the interview and screen-sharing data reported in Sections 3 through 6, and the design implications in Section 7 are explicitly framed as recommendations informed by those data rather than as derived conclusions. The lack of exposure-normalized counts in Table 3 and the missing codebook are methodological limitations that belong under validity or correctness risk, not circularity. No specific reduction of a claim to its own input can be quoted from the paper, so per the review rules the appropriate finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

The paper's qualitative claims rest on standard domain assumptions of interview research: representative stimulus selection, self-reported behavior and vision status, and valid coding. No free parameters or invented entities are introduced.

assumptions (4)
  • domain assumption The 12 pre-selected media items capture a representative range of AI-generated content across platforms, formats, and categories.
    Section 3.2 states the curation goal ('representative coverage of widely circulated AI-generated media'), but no external benchmark or user validation shows that these items are representative of the broader AIG ecosystem users encounter.
  • domain assumption Participant self-reports and in-session interactions reflect routine real-world behavior.
    Section 3.1 uses remote screen-sharing and interviewer nudging; the paper does not triangulate with diary or logging data to confirm that observed behavior matches everyday practice.
  • domain assumption The deductive codebook and the 0.80 inter-coder reliability yield valid thematic interpretations.
    Section 3.4 describes the coding process, but the codebook itself is missing from the manuscript ('Table??'), so the coding definitions cannot be inspected.
  • domain assumption Participants' sensory abilities are accurately categorized by self-reported vision status.
    Section 3.3 relies on self-described vision status (e.g., 'some light perception', 'totally blind') without an independent clinical assessment.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Signals of Provenance: Practices & Challenges of Navigating Indicators in AI-Generated Media for Sighted and Blind Individuals." pith.science (2026). https://pith.science/paper/DXLH6DXU

@misc{pith2026250516057,
  author       = {Pith},
  title        = {Pith review of: Signals of Provenance: Practices & Challenges of Navigating Indicators in AI-Generated Media for Sighted and Blind Individuals},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/DXLH6DXU}},
  note         = {Machine review of arXiv:2505.16057}
}
read the original abstract

AI-Generated (AIG) content has become increasingly widespread by recent advances in generative models and the easy-to-use tools that have significantly lowered the technical barriers for producing highly realistic audio, images, and videos through simple natural language prompts. In response, platforms are adopting provable provenance with platforms recommending AIG to be self-disclosed and signaled to users. However, these indicators may be often missed, especially when they rely solely on visual cues and make them ineffective to users with different sensory abilities. To address the gap, we conducted semi-structured interviews (N=28) with 15 sighted and 13 BLV participants to examine their interaction with AIG content through self-disclosed AI indicators. Our findings reveal diverse mental models and practices, highlighting different strengths and weaknesses of content-based (e.g., title, description) and menu-aided (e.g., AI labels) indicators. While sighted participants leveraged visual and audio cues, BLV participants primarily relied on audio and existing assistive tools, limiting their ability to identify AIG. Across both groups, they frequently overlooked menu-aided indicators deployed by platforms and rather interacted with content-based indicators such as title and comments. We uncovered usability challenges stemming from inconsistent indicator placement, unclear metadata, and cognitive overload. These issues were especially critical for BLV individuals due to the insufficient accessibility of interface elements. We provide practical recommendations and design implications for future AIG indicators across several dimensions.

Figures

Figures reproduced from arXiv: 2505.16057 by the authors.

Figure 1
Figure 1. User Interface of YouTube and TikTok’s AI info label. (a) single-line label "Altered or synthetic content" and hidden description "How this content was [PITH_FULL_IMAGE:figures/full_fig_p005_1.png] view at source ↗
Figure 2
Figure 2. User Interface of Meta’s AI info label. "AI info" label is embedded in the post header, rotating with other elements like the location tag. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Participants’ Mental Model for AI-generated Content - four dimensions derived from our qualitative results. [PITH_FULL_IMAGE:figures/full_fig_p008_3.png] view at source ↗
Figures from the paper (6 more)
Figure 4
Figure 4. Figure 4: Example screenshots of title and comments that participants accessed during interviews. (Left) The title is positioned very close to the content, making [PITH_FULL_IMAGE:figures/full_fig_p009_4.png]
Figure 5
Figure 5. Figure 5: Example screenshots of descriptions and hashtags that participants accessed during interviews. (Left) Descriptions are placed in highly visible areas but [PITH_FULL_IMAGE:figures/full_fig_p010_5.png]
Figure 6
Figure 6. Figure 6: Example screenshots of watermark and overlay texts that participants accessed during interviews. (Left two) Watermarks are placed in the top left [PITH_FULL_IMAGE:figures/full_fig_p011_6.png]
Figure 7
Figure 7. Figure 7: Example screenshots of menu-aided AI labels that participants accessed during interviews. (Left) YouTube’s AI label "How this content was made" was [PITH_FULL_IMAGE:figures/full_fig_p012_7.png]
Figure 8
Figure 8. Figure 8: Example screenshots of interface with accessibility challenges. (Left) automatically generated alt texts describing visual components of the image; [PITH_FULL_IMAGE:figures/full_fig_p012_8.png]
Figure 9
Figure 9. Figure 9: Screen captures of the AI-generated media examples used in the study. [PITH_FULL_IMAGE:figures/full_fig_p022_9.png]

Discussion (0). Sign in to comment.

Forward citations

Cited by 1 Pith paper

Reviewed papers in the Pith corpus that reference this work. Sorted by Pith novelty score. Full citation record

  1. More Human or More AI? Visualizing Human-AI Collaboration Disclosures in Journalistic News Production

    cs.HC 2026-01 conditional novelty 6.0 of 10

    Disclosure visualization format systematically shifts readers' perceptions of human vs AI contribution: role-based timelines amplify perceived AI role in mostly human articles, while task-based timelines make mostly A...

Reference graph

Works this paper leans on

138 extracted references · 53 canonical work pages · cited by 1 Pith paper

  1. [1]

    [n. d.]. About Generated Content — support.tiktok.com. https://support.tikt ok.com/en/using-tiktok/creating-videos/ai-generated-content. [Accessed 04-05-2025]

    2025

  2. [2]

    [n. d.]. Build trust with content credentials in Microsoft Designer | Learn at Microsoft Create — create.microsoft.com. https://create.microsoft.com/en- us/learn/articles/designer-content-credentials. [Accessed 19-01-2025]

    2025

  3. [3]

    [n. d.]. Content Credentials overview. https://helpx.adobe.com/creative- cloud/help/content-credentials.html. [Accessed 13-05-2025]

    2025

  4. [4]

    [n. d.]. Disclosing use of altered or synthetic content - Android - YouTube Help — support.google.com. https://support.google.com/youtube/answer/14328491?h l=en. [Accessed 05-05-2025]

    work page arXiv 2025

  5. [5]

    [n. d.]. EU AI Act: first regulation on artificial intelligence | Topics | European Parliament — europarl.europa.eu. https://www.europarl.europa.eu/topics/en/a rticle/20230601STO93804/eu-ai-act-first-regulation-on-artificial-intelligence. [Accessed 13-05-2025]

    work page arXiv 2025

  6. [6]

    [n. d.]. Help Center — help.instagram.com. https://help.instagram.com/7611219 59519495/?cms_platform=mobile-basic&helpref=platform_switcher. [Accessed 05-05-2025]

    work page arXiv 2025

  7. [7]

    [n. d.]. Introducing Official Content Credentials Icon - C2PA — c2pa.org. https: //c2pa.org/post/contentcredentials/. [Accessed 03-05-2025]

    2025

  8. [8]

    [n. d.]. Our approach to responsible AI innovation — blog.youtube. https: //blog.youtube/inside-youtube/our-approach-to-responsible-ai-innovation/. [Accessed 13-05-2025]

    2025

Show all 138 references
  1. [9]

    [n. d.]. Scammers use AI to mimic voices of loved ones in distress — cbsnews.com. https://www.cbsnews.com/news/scammers-ai-mimic-voices-loved-ones-in- distress/. [Accessed 03-05-2025]

    2025

  2. [10]

    deepfakes

    [n. d.]. Voters: Here’s how to spot AI “deepfakes” that spread election-related misinformation — heinz.cmu.edu. https://www.heinz.cmu.edu/media/2024/O ctober/voters-heres-how-to-spot-ai-deepfakes-that-spread-election-related- misinformation1. [Accessed 30-03-2025]

    2024

  3. [11]

    [n. d.]. ‘Deepfake’ of Biden’s Voice Called Early Example of US Election Dis- information — learningenglish.voanews.com. https://learningenglish.voanew s.com/a/deepfake-of-biden-s-voice-called-early-example-of-us-election- disinformation/7455392.html. [Accessed 30-03-2025]

    2025

  4. [12]

    New labels for disclosing AI-generated content - Newsroom | TikTok — newsroom.tiktok.com

    2023. New labels for disclosing AI-generated content - Newsroom | TikTok — newsroom.tiktok.com. https://newsroom.tiktok.com/en-us/new-labels-for- disclosing-ai-generated-content. [Accessed 05-05-2025]

    2023

  5. [13]

    How we’re helping creators disclose altered or synthetic content — blog.youtube

    2024. How we’re helping creators disclose altered or synthetic content — blog.youtube. https://blog.youtube/news-and-events/disclosing-ai-generated- content/. [Accessed 05-05-2025]

    2024

  6. [14]

    Ali Abdolrahmani and Ravi Kuber. 2016. Should I trust it when I cannot see it? Credibility assessment for blind web users. InProceedings of the 18th international acm sigaccess conference on computers and accessibility . 191–199. , Vol. 1, No. 1, Article . Publication date: Ma...

    2016

  7. [15]

    Dustin Adams, Lourdes Morales, and Sri Kurniawan. 2013. A qualitative study to support a blind photography mobile application. In Proceedings of the 6th Inter- national Conference on PErvasive Technologies Related to Assistive Environments . 1–8

    2013

  8. [16]

    Darius Afchar, Vincent Nozick, Junichi Yamagishi, and Isao Echizen. 2018. Mesonet: a compact facial video forgery detection network. In 2018 IEEE in- ternational workshop on information forensics and security (WIFS) . IEEE, 1–7

    2018

  9. [17]

    Sami Alanazi, Seemal Asif, Antoinette Caird-daley, and Irene Moulitsas. 2025. Unmasking deepfakes: a multidisciplinary examination of social impacts and regulatory responses. Human-Intelligent Systems Integration (2025), 1–23

    2025

  10. [18]

    Zaynab Almutairi and Hebah Elgibreen. 2022. A review of modern audio deepfake detection methods: challenges and future directions. Algorithms 15, 5 (2022), 155

    2022

  11. [19]

    Aboubakr Aqle, Kamran Khowaja, and Dena Al-Thani. 2020. Preliminary eval- uation of interactive search engine interface for visually impaired users. IEEE access 8 (2020), 45061–45070

    2020

  12. [20]

    Sercan Arik, Jitong Chen, Kainan Peng, Wei Ping, and Yanqi Zhou. 2018. Neural voice cloning with a few samples. Advances in neural information processing systems 31 (2018)

    2018

  13. [21]

    Muskan Arora, Kaushal Kishore Mishra, Mandeep Singh, Praveen Singh, and Rashmi Tripathi. 2024. Deepfake Technology and Its Implications for Influencer Marketing. In Navigating the World of Deepfake Technology . IGI Global, 66–90

    2024

  14. [22]

    Vian Bakir, Alexander Laffer, Andrew McStay, Diana Miranda, and Lachlan Urquhart. 2024. On manipulation by emotional AI: UK adults’ views and gover- nance implications. Frontiers in Sociology 9 (2024), 1339834

    2024

  15. [23]

    Soubhik Barari, Christopher Lucas, Kevin Munger, et al. 2021. Political deepfake videos misinform the public, but no more than other fake media. OSF Preprints 13 (2021), 1–16

    2021

  16. [24]

    Sarah Barrington, Emily A Cooper, and Hany Farid. 2025. People are poorly equipped to detect AI-powered voice clones.Scientific Reports 15, 1 (2025), 11004

    2025

  17. [25]

    Cynthia L Bennett, Jane E, Martez E Mott, Edward Cutrell, and Meredith Ringel Morris. 2018. How teens with visual impairments take, edit, and share photos on social media. In Proceedings of the 2018 CHI conference on human factors in computing systems. 1–12

    2018

  18. [26]

    Monika Bickert. 2024. Our Approach to Labeling AI-Generated Content and Manipulated Media — about.fb.com. https://about.fb.com/news/2024/04/metas- approach- to- labeling- ai- generated- content- and- manipulated- media/. [Accessed 05-05-2025]

    2024

  19. [27]

    Kelly Burke. 2024. Music sector workers to lose nearly a quarter of income to AI in next four years, global study finds — theguardian.com. https://www.thegua rdian.com/music/2024/dec/04/artificial-intelligence-music-industry-impact- income-loss. [Accessed 12-05-2025]

    2024

  20. [28]

    Brian Chen and Gregory W Wornell. 2001. Quantization index modulation: A class of provably good methods for digital watermarking and information embedding. IEEE Transactions on Information theory 47, 4 (2001), 1423–1443

    2001

  21. [29]

    Bobby Chesney and Danielle Citron. 2019. Deep fakes: A looming challenge for privacy, democracy, and national security. Calif. L. Rev. 107 (2019), 1753

    2019

  22. [30]

    Beomsang Cho, Binh M Le, Jiwon Kim, Simon Woo, Shahroz Tariq, Alsharif Abuadbba, and Kristen Moore. 2023. Towards understanding of deepfake videos in the wild. In Proceedings of the 32nd ACM International Conference on Informa- tion and Knowledge Management . 4530–4537

    2023

  23. [31]

    John Collomosse and Andy Parsons. 2024. To Authenticity, and Beyond! Building safe and fair generative AI upon the three pillars of provenance. IEEE Computer Graphics and Applications 44, 3 (2024), 82–90

    2024

  24. [32]

    Ingemar J Cox, Joe Kilian, F Thomson Leighton, and Talal Shamoon. 1997. Secure spread spectrum watermarking for multimedia. IEEE transactions on image processing 6, 12 (1997), 1673–1687

    1997

  25. [33]

    Maitraye Das, Alexander J Fiannaca, Meredith Ringel Morris, Shaun K Kane, and Cynthia L Bennett. 2024. From provenance to aberrations: Image creator and screen reader user perspectives on alt text for AI-generated images. In Proceedings of the 2024 CHI Conference on Human Fact...

    2024

  26. [34]

    Trisha Datta, Binyi Chen, and Dan Boneh. 2024. VerITAS: verifying image transformations at scale. Cryptology ePrint Archive (2024)

    2024

  27. [35]

    Abhishek Dixit, Nirmal Kaur, and Staffy Kingra. 2023. Review of audio deepfake detection techniques: Issues and prospects. Expert Systems 40, 8 (2023), e13322

    2023

  28. [36]

    Brian Dolhansky, Joanna Bitton, Ben Pflaum, Jikuo Lu, Russ Howes, Menglin Wang, and Cristian Canton Ferrer. 2020. The deepfake detection challenge (dfdc) dataset. arXiv preprint arXiv:2006.07397 (2020)

    2020 arXiv

  29. [37]

    Paul England, Henrique S Malvar, Eric Horvitz, Jack W Stokes, Cédric Fournet, Rebecca Burke-Aguero, Amaury Chamayou, Sylvan Clebsch, Manuel Costa, John Deutscher, et al. 2021. AMP: Authentication of media via provenance. In Proceedings of the 12th ACM Multimedia Systems Confer...

    2021

  30. [38]

    Jason K Eshraghian. 2020. Human ownership of artificial creativity. Nature Machine Intelligence 2, 3 (2020), 157–160

    2020

  31. [39]

    Hany Farid. 2022. Creating, using, misusing, and detecting deep fakes. Journal of Online Trust and Safety 1, 4 (2022)

    2022

  32. [40]

    KJ Kevin Feng, Nick Ritchie, Pia Blumenthal, Andy Parsons, and Amy X Zhang

  33. [41]

    Jennifer Fereday and Eimear Muir-Cochrane. 2006. Demonstrating rigor using thematic analysis: A hybrid approach of inductive and deductive coding and theme development. International journal of qualitative methods 5, 1 (2006), 80–92

    2006

  34. [42]

    Coalition for Content Provenance and Authenticity. [n. d.]. Guiding Principles for C2PA Designs and Specifications. Accessed on 12.11.2024. https://c2pa.org/p rinciples/

    2024

  35. [43]

    Coaliton for Content Provenance and Authenticity. 2025. Content Credentials — contentcredentials.org. https://contentcredentials.org/. [Accessed 02-05-2025]

    2025

  36. [44]

    Coaliton for Content Provenance and Authenticity. 2025. Content Credentials : C2PA Technical Specification :: C2PA Specifications — c2pa.org. https://c2pa.o rg/specifications/specifications/2.1/specs/C2PA_Specification.html. [Accessed 02-05-2025]

    2025

  37. [45]

    Tao Fu, Ming Xia, and Gaobo Yang. 2022. Detecting GAN-generated face im- ages via hybrid texture and sensor noise based features. Multimedia Tools and Applications 81, 18 (2022), 26345–26359

    2022

  38. [46]

    Dilrukshi Gamage, Dilki Sewwandi, Min Zhang, and Arosha Bandara. 2025. Labeling Synthetic Content: User Perceptions of Warning Label Designs for AI-generated Content on Social Media. arXiv preprint arXiv:2503.05711 (2025)

    2025 arXiv

  39. [47]

    Mingkun Gao, Ziang Xiao, Karrie Karahalios, and Wai-Tat Fu. 2018. To label or not to label: The effect of stance and credibility labels on readers’ selection and perception of news articles. Proceedings of the ACM on Human-Computer Interaction 2, CSCW (2018), 1–16

    2018

  40. [48]

    Ricardo E Gonzalez Penuela, Paul Vermette, Zihan Yan, Cheng Zhang, Keith Vertanen, and Shiri Azenkot. 2022. Understanding How People with Visual Impairments Take Selfies: Experiences and Challenges. In Proceedings of the 24th International ACM SIGACCESS Conference on Computers...

    2022

  41. [49]

    Ian Goodfellow, Jean Pouget-Abadie, Mehdi Mirza, Bing Xu, David Warde-Farley, Sherjil Ozair, Aaron Courville, and Yoshua Bengio. 2020. Generative adversarial networks. Commun. ACM 63, 11 (2020), 139–144

    2020

  42. [50]

    Google. 2025. New Disclosures and Labels for Generative AI Content on YouTube - YouTube Community — support.google.com. https://support.google.com/you tube/thread/264550152/new-disclosures-and-labels-for-generative-ai-content- on-youtube. [Accessed 15-03-2025]

    2025

  43. [51]

    Matthew Groh, Ziv Epstein, Chaz Firestone, and Rosalind Picard. 2022. Deep- fake detection by human crowds, machines, and machine-informed crowds. Proceedings of the National Academy of Sciences 119, 1 (2022), e2110013119

    2022

  44. [52]

    David Güera and Edward J Delp. 2018. Deepfake video detection using recurrent neural networks. In 2018 15th IEEE international conference on advanced video and signal based surveillance (A VSS). IEEE, 1–6

    2018

  45. [53]

    Sam Gunn, Xuandong Zhao, and Dawn Song. 2024. An undetectable watermark for generative image models. arXiv preprint arXiv:2410.07369 (2024)

    2024 arXiv

  46. [54]

    Ameer Hamza, Abdul Rehman Rehman Javed, Farkhund Iqbal, Natalia Kryvinska, Ahmad S Almadhor, Zunera Jalil, and Rouba Borghol. 2022. Deepfake audio detection via MFCC features using machine learning. IEEE Access 10 (2022), 134018–134028

    2022

  47. [55]

    Chaeeun Han, Prasenjit Mitra, and Syed Masum Billah. 2024. Uncovering human traits in determining real and spoofed audio: Insights from blind and sighted individuals. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems. 1–14

    2024

  48. [56]

    Yuanning Han, Ziyi Qiu, Jiale Cheng, and Ray Lc. 2024. When teams embrace AI: human collaboration strategies in generative prompting in a creative design task. In Proceedings of the 2024 CHI Conference on Human Factors in Computing Systems. 1–14

    2024

  49. [57]

    Susumu Harada, Daisuke Sato, Dustin W Adams, Sri Kurniawan, Hironobu Takagi, and Chieko Asakawa. 2013. Accessible photo album: enhancing the photo sharing experience for people with visual impairment. In Proceedings of the SIGCHI conference on human factors in computing system...

    2013

  50. [58]

    Ammarah Hashmi, Sahibzada Adil Shahzad, Chia-Wen Lin, Yu Tsao, and Hsin- Min Wang. 2024. Unmasking illusions: Understanding human perception of audiovisual deepfakes. arXiv preprint arXiv:2405.04097 (2024)

    2024 arXiv

  51. [59]

    Runyi Hu, Jie Zhang, Yiming Li, Jiwei Li, Qing Guo, Han Qiu, and Tianwei Zhang

  52. [60]

    Xuming Hu, Hanqian Li, Jungang Li, and Aiwei Liu. 2025. VideoMark: A Distortion-Free Robust Watermarking Framework for Video Diffusion Mod- els. arXiv preprint arXiv:2504.16359 (2025)

    2025

  53. [61]

    Yiqing Hua, Shuo Niu, Jie Cai, Lydia B Chilton, Hendrik Heuer, and Donghee Yvette Wohn. 2024. Generative AI in user-generated content. In Ex- tended Abstracts of the CHI Conference on Human Factors in Computing Systems . , Vol. 1, No. 1, Article . Publication date: May 2025. 1...

    2024

  54. [62]

    Ayae Ide and Tanusree Sharma. 2025. Personhood Credentials: Human-Centered Design Recommendation Balancing Security, Usability, and Trust. arXiv preprint arXiv:2502.16375 (2025)

    2025 arXiv

  55. [63]

    Content Authenticity Initiative. 2025. Content Authenticity Initiative — con- tentauthenticity.org. https://contentauthenticity.org/. [Accessed 02-05-2025]

    2025

  56. [64]

    Chandrika Jayant, Hanjie Ji, Samuel White, and Jeffrey P Bigham. 2011. Sup- porting blind photography. In The proceedings of the 13th international ACM SIGACCESS conference on Computers and accessibility . 203–210

    2011

  57. [65]

    Yan Ju, Shu Hu, Shan Jia, George H Chen, and Siwei Lyu. 2024. Improving fairness in deepfake detection. In Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision . 4655–4665

    2024

  58. [66]

    2023.{GuardLens}: Supporting safer online browsing for people with visual impairments

    Smirity Kaushik, Natã M Barbosa, Yaman Yu, Tanusree Sharma, Zachary Kilhoffer, JooYoung Seo, Sauvik Das, and Yang Wang. 2023.{GuardLens}: Supporting safer online browsing for people with visual impairments. In Nineteenth Symposium on Usable Privacy and Security (SOUPS 2023) . 361–380

    2023

  59. [67]

    Eric Kee, Micah K Johnson, and Hany Farid. 2011. Digital image authentication from JPEG headers. IEEE transactions on information forensics and security 6, 3 (2011), 1066–1075

    2011

  60. [68]

    Prakash L Kharvi. 2024. Understanding the Impact of AI-Generated Deepfakes on Public Opinion, Political Discourse, and Personal Security in Social Media. IEEE Security & Privacy (2024)

    2024

  61. [69]

    Diederik P Kingma, Max Welling, et al. 2013. Auto-encoding variational bayes

    2013

  62. [70]

    Nils C Köbis, Barbora Doležalová, and Ivan Soraperra. 2021. Fooled twice: People cannot detect deepfakes but think they can. Iscience 24, 11 (2021)

    2021

  63. [71]

    Seb Joseph Krystal Scanlon. 2024. TikTok joins the AI-driven advertising pack to compete with Meta for ad dollars — digiday.com. https://digiday.com/market ing/tiktok-joins-the-ai-driven-advertising-pack-to-compete-with-meta-for- ad-dollars/. [Accessed 12-05-2025]

    2024

  64. [72]

    Lin Kyi, Amruta Mahuli, M Silberman, Reuben Binns, Jun Zhao, and Asia J Biega. 2025. Governance of Generative AI in Creative Work: Consent, Credit, Compensation, and Beyond. arXiv preprint arXiv:2501.11457 (2025)

    2025 arXiv

  65. [73]

    Linda Laurier, Ave Giulietta, Arlo Octavia, and Meade Cleti. 2024. The Cat and Mouse Game: The Ongoing Arms Race Between Diffusion Models and Detection Methods. arXiv preprint arXiv:2410.18866 (2024)

    2024 arXiv

  66. [74]

    Adrian Lecaros, Freddy Paz, and Arturo Moquillaza. 2021. Challenges and opportunities on the application of heuristic evaluations: a systematic literature review. In International Conference on Human-Computer Interaction . Springer, 242–261

    2021

  67. [75]

    Chang-Tsun Li, Karunakar A Kotegar, et al. 2025. AI-Synthesized Image De- tection: Source Camera Fingerprinting to Discern the Authenticity of Digital Images. IEEE Access (2025)

    2025

  68. [76]

    Yuezun Li, Ming-Ching Chang, and Siwei Lyu. 2018. In ictu oculi: Exposing ai created fake videos by detecting eye blinking. In2018 IEEE International workshop on information forensics and security (WIFS) . Ieee, 1–7

    2018

  69. [77]

    Gionnieve Lim and Simon T Perrault. 2023. Effects of automated misinformation warning labels on the intents to like, comment and share posts. In Proceedings of the 11th International Conference on Human-Agent Interaction . 299–305

    2023

  70. [78]

    Chi Liu, Tianqing Zhu, Yuan Zhao, Jun Zhang, and Wanlei Zhou. 2024. Disentan- gling different levels of GAN fingerprints for task-specific forensics. Computer Standards & Interfaces 89 (2024), 103825

    2024

  71. [79]

    Juniper Lovato, Julia Witte Zimmerman, Isabelle Smith, Peter Dodds, and Jen- nifer L Karson. 2024. Foregrounding artist opinions: A survey study on trans- parency, ownership, and fairness in AI generative art. In Proceedings of the aaai/acm conference on ai, ethics, and societ...

    2024

  72. [80]

    Shilin Lu, Zihan Zhou, Jiayou Lu, Yuanzhi Zhu, and Adams Wai-Kin Kong

  73. [81]

    Haley MacLeod, Cynthia L Bennett, Meredith Ringel Morris, and Edward Cutrell

  74. [82]

    Kimberly T Mai, Sergi Bray, Toby Davies, and Lewis D Griffin. 2023. Warning: Humans cannot reliably detect speech deepfakes. Plos one 18, 8 (2023), e0285333

    2023

  75. [83]

    Mehtab Malik. 2023. Virtual influencers are earning as much as their human counterparts — wired.me. https://wired.me/culture/virtual- influencer/. [Accessed 12-05-2025]

    2023

  76. [84]

    Karolina Mania. 2024. Legal protection of revenge and deepfake porn victims in the European Union: Findings from a comparative legal study. Trauma, Violence, & Abuse 25, 1 (2024), 117–129

    2024

  77. [85]

    Marie-Helen Maras and Alex Alexandrou. 2019. Determining authenticity of video evidence in the age of artificial intelligence and in the wake of Deepfake videos. The international journal of evidence & proof 23, 3 (2019), 255–262

    2019

  78. [86]

    Mary L McHugh. 2012. Interrater reliability: the kappa statistic. Biochemia medica 22, 3 (2012), 276–282

    2012

  79. [87]

    Ivan Mehta. 2025. AvatarOS snags $7M seed round from M13 to build an AI- powered virtual influencer platform | TechCrunch — techcrunch.com. https: //techcrunch.com/2025/03/10/avataros-snags-7m-seed-round-from-m13-to- build-an-ai-powered-virtual-influencer-platform/. [Accessed ...

    2025

  80. [88]

    Meta. 2025. Labeling AI-Generated Images on Facebook, Instagram and Threads | Meta — about.fb.com. https://about.fb.com/news/2024/02/labeling-ai-generated- images-on-facebook-instagram-and-threads/. [Accessed 15-03-2025]

    2025

  81. [89]

    Jaron Mink, Licheng Luo, Natã M Barbosa, Olivia Figueira, Yang Wang, and Gang Wang. 2022. {DeepPhish}: Understanding user trust towards artificially generated profiles in online social networks. In 31st USENIX Security Symposium (USENIX Security 22). 1669–1686

    2022

  82. [90]

    Jaron Mink, Miranda Wei, Collins W Munyendo, Kurt Hugenberg, Tadayoshi Kohno, Elissa M Redmiles, and Gang Wang. 2024. It’s Trying Too Hard To Look Real: Deepfake Moderation Mistakes and Identity-Based Bias. In Proceedings of the 2024 CHI Conference on Human Factors in Computin...

    2024

  83. [91]

    Meredith Ringel Morris and Jed R Brubaker. 2024. Generative ghosts: Anticipating benefits and risks of AI afterlives. arXiv preprint arXiv:2402.01662 (2024)

    2024 arXiv

  84. [92]

    Rami Mubarak, Tariq Alsboui, Omar Alshaikh, Isa Inuwa-Dutse, Saad Khan, and Simon Parkinson. 2023. A survey on the detection and impacts of deepfakes in visual, audio, and textual formats. Ieee Access 11 (2023), 144497–144529

    2023

  85. [93]

    Nicolas M Müller, Karla Pizzi, and Jennifer Williams. 2022. Human perception of audio deepfakes. In Proceedings of the 1st international workshop on deepfake detection for audio multimedia . 85–91

    2022

  86. [94]

    ABC News. 2025. Catholic community reacts to Trump’s AI image of himself as the pope — abcnews.go.com. https://abcnews.go.com/Politics/catholic- community-reacts-trumps-ai-image-pope/story?id=121447607. [Accessed 12-05-2025]

    2025

  87. [95]

    Sophie J Nightingale and Hany Farid. 2022. AI-synthesized faces are indistin- guishable from real faces and more trustworthy. Proceedings of the National Academy of Sciences 119, 8 (2022), e2120481119

    2022

  88. [96]

    I’m a Solo Developer but AI is My New Ill-Informed Co-Worker

    Ruchi Panchanadikar and Guo Freeman. 2024. " I’m a Solo Developer but AI is My New Ill-Informed Co-Worker": Envisioning and Designing Generative AI to Support Indie Game Development. Proceedings of the ACM on Human-Computer Interaction 8, CHI PLAY (2024), 1–26

    2024

  89. [97]

    Sohyun Park, Someen Park, Jaehoon Kim, and Kyungsik Han. 2024. Exploring the Impact of AI-Generated Images on Political News Perception and Under- standing. In Companion Publication of the 2024 Conference on Computer-Supported Cooperative Work and Social Computing . 565–571

    2024

  90. [98]

    Andy Parsons. [n. d.]. Introducing Adobe Content Authenticity: A free web app to help creators protect their work, gain attribution and build trust | Adobe Blog — blog.adobe.com. https://blog.adobe.com/en/publish/2024/10/08/introducing- adobe-content-authenticity-free-web-app-...

    2024

  91. [99]

    Reza Arkan Partadiredja, Carlos Entrena Serrano, and Davor Ljubenkov. 2020. AI or human: the socio-ethical implications of AI-generated media content. In 2020 13th CMI Conference on Cybersecurity and Privacy (CMI)-Digital Transformation- Potentials and Challenges (51275) . IEEE, 1–6

    2020

  92. [100]

    Matías Pizarro, Mike Laszkiewicz, Dorothea Kolossa, and Asja Fischer. 2024. Single-Model Attribution for Spoofed Speech via Vocoder Fingerprints in an Open-World Setting. arXiv preprint arXiv:2411.14013 (2024)

    2024

  93. [101]

    Perils Promises. 2023. Labeling AI-Generated Content. (2023)

    2023

  94. [102]

    Andreas Rossler, Davide Cozzolino, Luisa Verdoliva, Christian Riess, Justus Thies, and Matthias Nießner. 2019. Faceforensics++: Learning to detect manipulated facial images. In Proceedings of the IEEE/CVF international conference on computer vision. 1–11

    2019

  95. [103]

    Emily Saltz, Claire R Leibowicz, and Claire Wardle. 2021. Encounters with visual misinformation and labels across platforms: An interview and diary study to inform ecosystem approaches to misinformation interventions. In Extended Abstracts of the 2021 CHI Conference on Human F...

    2021

  96. [104]

    Pamela Samuelson. 2023. Generative AI meets copyright. Science 381, 6654 (2023), 158–161

    2023

  97. [105]

    Marc Schneider and Shih-Fu Chang. 1996. A robust content based digital signa- ture for image authentication. In Proceedings of 3rd IEEE international conference on image processing, Vol. 3. IEEE, 227–230

    1996

  98. [106]

    Letícia Seixas Pereira, José Coelho, André Rodrigues, João Guerreiro, Tiago Guerreiro, and Carlos Duarte. 2022. Authoring accessible media content on social networks. In Proceedings of the 24th International ACM SIGACCESS Conference on Computers and Accessibility . 1–11

    2022

  99. [107]

    Shawn Shan, Jenna Cryan, Emily Wenger, Haitao Zheng, Rana Hanocka, and Ben Y Zhao. 2023. Glaze: Protecting artists from style mimicry by {Text-to- Image} models. In 32nd USENIX Security Symposium (USENIX Security 23) . 2187– 2204

    2023

  100. [108]

    I Just Didn’t Notice It:

    Filipo Sharevski and Aziz N Zeidieh. 2023. “I Just Didn’t Notice It:” Experiences with Misinformation Warnings on Social Media amongst Users Who Are Low Vision or Blind. In Proceedings of the 2023 New Security Paradigms Workshop . , Vol. 1, No. 1, Article . Publication date: M...

    2023

  101. [109]

    I’m not con- vinced that they don’t collect more than is necessary

    Tanusree Sharma, Lin Kyi, Yang Wang, and Asia J Biega. 2024. " I’m not con- vinced that they don’t collect more than is necessary":{User-Controlled} Data Minimization Design in Search Engines. In 33rd USENIX Security Symposium (USENIX Security 24). 2797–2812

    2024

  102. [110]

    Tanusree Sharma, Yujin Potter, Zachary Kilhoffer, Yun Huang, Dawn Song, and Yang Wang. 2024. From Experts to the Public: Governing Multimodal Language Models in Politically Sensitive Video Analysis. arXiv preprint arXiv:2410.01817 (2024)

    2024 arXiv

  103. [111]

    Tanusree Sharma, Abigale Stangl, Lotus Zhang, Yu-Yun Tseng, Inan Xu, Leah Findlater, Danna Gurari, and Yang Wang. 2023. Disability-first design and creation of a dataset showing private visual information collected with people who are blind. In Proceedings of the 2023 CHI Conf...

    2023

  104. [112]

    Cuihua Shen, Mona Kasra, and James O’Brien. 2021. This photograph has been altered: Testing the effectiveness of image forensic labeling on news image credibility. arXiv preprint arXiv:2101.07951 (2021)

    2021 arXiv

  105. [113]

    Nyein Nyein Thaw, Thin July, Aye Nu Wai, Dion Hoe-Lian Goh, and Alton YK Chua. 2021. How are deepfake videos detected? An initial user study. In HCI International 2021-Posters: 23rd HCI International Conference, HCII 2021, Virtual Event, July 24–29, 2021, Proceedings, Part I 2...

    2021

  106. [114]

    TikTok. 2025. About AI-generated content | TikTok Help Center — sup- port.tiktok.com. https://support.tiktok.com/en/using- tiktok/creating- videos/ai-generated-content. [Accessed 15-03-2025]

    2025

  107. [115]

    Maddalena Torricelli, Mauro Martino, Andrea Baronchelli, and Luca Maria Aiello

  108. [116]

    Yu-Yun Tseng, Tanusree Sharma, Lotus Zhang, Abigale Stangl, Leah Findlater, Yang Wang, and Danna Gurari. 2025. Biv-priv-seg: Locating private content in images taken by people with visual impairments. In 2025 IEEE/CVF Winter Conference on Applications of Computer Vision (W ACV...

    2025

  109. [117]

    Cristian Vaccari and Andrew Chadwick. 2020. Deepfakes and disinformation: Exploring the impact of synthetic political video on deception, uncertainty, and trust in news. Social media+ society 6, 1 (2020), 2056305120903408

    2020

  110. [118]

    Luisa Verdoliva. 2020. Media forensics and deepfakes: an overview. IEEE journal of selected topics in signal processing 14, 5 (2020), 910–932

    2020

  111. [119]

    Violeta Voykinska, Shiri Azenkot, Shaomei Wu, and Gilly Leshed. 2016. How blind people interact with visual content on social networking services. In Pro- ceedings of the 19th acm conference on computer-supported cooperative work & social computing. 1584–1595

    2016

  112. [120]

    In Proceedings of the 16th ACM Web Science Conference

    The role of interface design on prompt-mediated creativity in Generative AI. In Proceedings of the 16th ACM Web Science Conference . 235–240

  113. [121]

    Yuxin Wen, John Kirchenbauer, Jonas Geiping, and Tom Goldstein. 2023. Tree- ring watermarks: Fingerprints for diffusion images that are invisible and robust. arXiv preprint arXiv:2305.20030 (2023)

    2023 arXiv

  114. [122]

    Yuxin Xu, Mengqiu Cheng, and Anastasia Kuzminykh. 2024. What Makes It Mine? Exploring Psychological Ownership over Human-AI Co-Creations. In Proceedings of the 50th Graphics Interface Conference . 1–8

    2024

  115. [123]

    Xinrui Yan, Jiangyan Yi, Jianhua Tao, Chenglong Wang, Haoxin Ma, Tao Wang, Shiming Wang, and Ruibo Fu. 2022. An initial investigation for detecting vocoder fingerprints of fake audio. In Proceedings of the 1st International Workshop on Deepfake Detection for Audio Multimedia . 61–68

    2022

  116. [124]

    Xuezi Dan Yan Luo. [n. d.]. China Releases New Labeling Requirements for AI-Generated Content — insideprivacy.com. https://www.insideprivacy.co m/international/china/china-releases-new-labeling-requirements-for-ai- generated-content/. [Accessed 13-05-2025]

    2025

  117. [125]

    Better Be Computer or I’m Dumb

    Kevin Warren, Tyler Tucker, Anna Crowder, Daniel Olszewski, Allison Lu, Car- oline Fedele, Magdalena Pasternak, Seth Layton, Kevin Butler, Carrie Gates, et al. 2024. " Better Be Computer or I’m Dumb": A Large-Scale Evaluation of Humans as Audio Deepfake Detectors. InProceeding...

    2024

  118. [126]

    Michael Yankoski, Walter Scheirer, and Tim Weninger. 2021. Meme warfare: AI countermeasures to disinformation should focus on popular, not perfect, fakes. Bulletin of the atomic scientists 77, 3 (2021), 119–123

    2021

  119. [127]

    Ning Yu, Larry S Davis, and Mario Fritz. 2019. Attributing fake images to gans: Learning and analyzing gan fingerprints. In Proceedings of the IEEE/CVF international conference on computer vision . 7556–7566

    2019

  120. [128]

    Peipeng Yu, Zhihua Xia, Jianwei Fei, and Yujiang Lu. 2021. A survey on deepfake video detection. Iet Biometrics 10, 6 (2021), 607–624

    2021

  121. [129]

    Han Zhang, Tao Xu, Hongsheng Li, Shaoting Zhang, Xiaogang Wang, Xiaolei Huang, and Dimitris N Metaxas. 2017. Stackgan: Text to photo-realistic image synthesis with stacked generative adversarial networks. In Proceedings of the IEEE international conference on computer vision ....

    2017

  122. [130]

    Xin Yang, Yuezun Li, and Siwei Lyu. 2019. Exposing deep fakes using inconsistent head poses. In ICASSP 2019-2019 IEEE international conference on acoustics, speech and signal processing (ICASSP) . IEEE, 8261–8265

    2019

  123. [131]

    Yipin Zhou and Ser-Nam Lim. 2021. Joint audio-visual deepfake detection. In Proceedings of the IEEE/CVF International Conference on Computer Vision . 14800– 14809

    2021

  124. [132]

    Zhixuan Zhou, Tanusree Sharma, Luke Emano, Sauvik Das, and Yang Wang

  125. [135]

    Lotus Zhang, Abigale Stangl, Tanusree Sharma, Yu-Yun Tseng, Inan Xu, Danna Gurari, Yang Wang, and Leah Findlater. 2024. Designing Accessible Obfuscation Support for Blind Individuals’ Visual Privacy Management. In Proceedings of the 2024 CHI Conference on Human Factors in Comp...

    2024

  126. [138]

    AI-assisted

    Iterative design of an accessible crypto wallet for blind users. InNineteenth Symposium on Usable Privacy and Security (SOUPS 2023) . 381–398. , Vol. 1, No. 1, Article . Publication date: May 2025. 20 • Ayae Ide, Tory Park, Jaron Mink, and Tanusree Sharma Table 5. Participant ...

    2023

  127. [2017]

    In proceedings of the 2017 CHI conference on human factors in computing systems

    Understanding blind people’s experiences with computer-generated cap- tions of social media images. In proceedings of the 2017 CHI conference on human factors in computing systems . 5988–5999

    2017

  128. [2023]

    Proceedings of the ACM on Human-Computer Interaction 7, CSCW2 (2023), 1–42

    Examining the impact of provenance-enabled media on trust and accuracy perceptions. Proceedings of the ACM on Human-Computer Interaction 7, CSCW2 (2023), 1–42

    2023

  129. [2024]

    arXiv preprint arXiv:2410.18775 (2024)

    Robust watermarking using generative priors against image editing: From benchmarking to advances. arXiv preprint arXiv:2410.18775 (2024)

    2024 arXiv

  130. [2025]

    arXiv preprint arXiv:2501.14195 (2025)

    VideoShield: Regulating Diffusion-based Video Generation Models via Watermarking. arXiv preprint arXiv:2501.14195 (2025)

    2025 arXiv

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.