pith. sign in

arxiv: 2604.12190 · v1 · submitted 2026-04-14 · 💻 cs.CY · cs.AI· cs.HC

Characterizing Resource Sharing Practices on Underground Internet Forum Synthetic Non-Consensual Intimate Image Content Creation Communities

Bernardo B. P. Medeiros (1) , Malvika Jadhav (1) , Allison Lu (1) , Tadayoshi Kohno (2) , Vincent Bindschaedler (1) , Kevin R. B. Butler (1) ((1) University of Florida , (2) Georgetown University) This is my paper

Pith reviewed 2026-05-10 16:12 UTC · model grok-4.3

classification 💻 cs.CY cs.AIcs.HC
keywords SNCIIsynthetic non-consensual intimate imagesunderground forumsresource sharingknowledge transfer4chanredditcontent creation communities
0
0 comments X

The pith

Analysis of 4chan and Reddit shows users of all technical levels share resources for creating and spreading synthetic non-consensual intimate images.

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

The paper examines hundreds of thousands of posts from two major underground forums to map how participants with different expertise levels interact in SNCII communities. It establishes that both sophisticated and novice users exchange primary tools for generating such content along with secondary tools for distribution, and that experts transfer knowledge to beginners in ways that extend these practices. A sympathetic reader would care because this mapping identifies concrete points where current regulations fall short and where focused actions might limit the ecosystem's growth.

Core claim

Through an integrated analysis of 282,154 4chan comments and 78,308 Reddit submissions spanning 165 days, the study characterizes the actors, actions, and resources in SNCII content creation communities. The central finding is that users with differing levels of technical sophistication employ and share a wide range of primary resources facilitating SNCII content creation as well as numerous secondary resources facilitating dissemination, while knowledge transfer between experts and newcomers facilitates propagation of these illicit resources. This empirical basis leads directly to the identification of gaps in existing regulatory infrastructure and the synthesis of critical intervention for

What carries the argument

Integrated multi-community analysis of forum posts that distinguishes primary resources for content creation, secondary resources for dissemination, and the knowledge transfer process between users of varying technical sophistication.

If this is right

  • Regulatory infrastructure contains identifiable gaps that can be addressed by focusing on resource flows rather than only final content.
  • Distinguishing primary creation resources from secondary dissemination resources enables more precise monitoring and disruption strategies.
  • Knowledge transfer channels between experts and newcomers represent actionable targets for reducing propagation of illicit methods.
  • Deterrence efforts must account for users at different technical sophistication levels to be effective.
  • Multi-platform coordination is required because the ecosystem spans communities with distinct user bases.

Where Pith is reading between the lines

These are editorial extensions of the paper, not claims the author makes directly.

  • Disrupting the documented knowledge transfer process could slow the spread of creation techniques more effectively than removing individual images after they appear.
  • The same data-driven mapping of resource sharing could be applied to other categories of online illicit activity to reveal common patterns.
  • As generative AI tools become easier to use, the role of expert-to-novice knowledge sharing identified here is likely to increase rather than diminish.
  • Platform policies that monitor discussions of resources and tutorials, in addition to posted content, may yield earlier detection of emerging practices.

Load-bearing premise

The sampled 4chan comments and Reddit submissions from the 165-day period accurately capture the key actors, actions, and resources across the broader SNCII ecosystem without significant bias from platform dynamics or anonymity.

What would settle it

A follow-up collection of posts from additional forums or a longer time window that finds little or no evidence of primary and secondary resource sharing or expert-to-novice knowledge transfer would contradict the core observations.

Figures

Figures reproduced from arXiv: 2604.12190 by (2) Georgetown University), Allison Lu (1), Bernardo B. P. Medeiros (1), Kevin R. B. Butler (1) ((1) University of Florida, Malvika Jadhav (1), Tadayoshi Kohno (2), Vincent Bindschaedler (1).

Figure 1
Figure 1. Figure 1: Overview of our dual-pipeline research methodol [PITH_FULL_IMAGE:figures/full_fig_p003_1.png] view at source ↗
Figure 2
Figure 2. Figure 2: Primary and secondary ecosystem stakeholders. [PITH_FULL_IMAGE:figures/full_fig_p005_2.png] view at source ↗
Figure 3
Figure 3. Figure 3: [4chan] Resource mentions by month. A brief spike in July-August 2025 follows the removal of real-person like￾ness models from Civitai in May 2025 [7]. resource requests seeking tools for independent generation. A spike in activity only two months after Civitai’s removal of real-person likeness models further suggests possible dis￾placement of SNCII-related resource sharing to 4chan [7]. 5.1.2 Influential … view at source ↗
Figure 4
Figure 4. Figure 4: [4chan] Monthly generation request volume and fulfillment (June 9–November 21, 2025). Request rates are high but fulfillment is very low, suggesting users desire SNCII content but may not receive the generations they desire from others. Veterans. We define VETERANS as users who either have a username or engage in meta discussions about 4chan that demonstrate historical knowledge of the ecosystem. VETER￾ANS… view at source ↗
Figure 5
Figure 5. Figure 5: User U1 first requests guidance on training an AI [PITH_FULL_IMAGE:figures/full_fig_p008_5.png] view at source ↗
Figure 6
Figure 6. Figure 6: [Reddit] Monthly commercial nudification app men￾tions. Application A mentions dwindle following terms of service changes in October 2025. we analyze (see [PITH_FULL_IMAGE:figures/full_fig_p009_6.png] view at source ↗
read the original abstract

Many malicious actors responsible for disseminating synthetic non-consensual intimate imagery (SNCII) operate within internet forums to exchange resources, strategies, and generated content across multiple platforms. Technically-sophisticated actors gravitate toward certain communities (e.g., 4chan), while lower-sophistication end-users are more active on others (e.g., Reddit). To characterize key stakeholders in the broader ecosystem, we perform an integrated analysis of multiple communities, analyzing 282,154 4chan comments and 78,308 Reddit submissions spanning 165 days between June and November 2025 to characterize involved actors, actions, and resources. We find: (a) that users with differing levels of technical sophistication employ and share a wide range of primary resources facilitating SNCII content creation as well as numerous secondary resources facilitating dissemination; and (b) that knowledge transfer between experts and newcomers facilitates propagation of these illicit resources. Based on our empirical analysis, we identify gaps in existing SNCII regulatory infrastructure and synthesize several critical intervention points for bolstering deterrence.

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, simulated authors' rebuttal, and a circularity audit. Tearing a paper down is the easy half of reading it; the pith above is the substance, this is the friction.

Referee Report

1 major / 1 minor

Summary. The manuscript analyzes 282,154 4chan comments and 78,308 Reddit submissions collected over a 165-day window (June–November 2025) to characterize actors, actions, and resources in underground communities engaged in synthetic non-consensual intimate image (SNCII) creation and dissemination. It reports that users of differing technical sophistication share a range of primary resources for content generation and secondary resources for distribution, with expert-to-newcomer knowledge transfer enabling propagation of these resources, and identifies regulatory gaps plus potential intervention points.

Significance. If the empirical patterns hold after addressing sampling limitations, the work supplies large-scale observational data on resource-sharing dynamics across platforms with different user sophistication profiles. This could help map the SNCII ecosystem and guide targeted deterrence. The scale of the collected corpus is a clear asset for an observational study in this domain.

major comments (1)
  1. [Data collection and analysis sections] The central claims rest on the assumption that the sampled 4chan and Reddit posts accurately capture primary/secondary resources and expert-to-newcomer knowledge transfer across the broader SNCII ecosystem. No validation, cross-platform checks, or explicit discussion of platform-specific biases (anonymity on 4chan versus moderation on Reddit) or temporal limitations of the 165-day window is provided, leaving open the possibility that observed patterns are artifacts of sampling rather than general features.
minor comments (1)
  1. [Abstract] The abstract states high-level findings on resource types and knowledge transfer but supplies no indication of the concrete analytical procedures (e.g., how resources were classified, how technical sophistication was operationalized, or inter-rater reliability for coding).

Simulated Author's Rebuttal

1 responses · 0 unresolved

We thank the referee for the detailed feedback, which highlights important considerations for strengthening the presentation of our observational study. We agree that the manuscript would benefit from more explicit discussion of sampling limitations, platform biases, and the scope of our claims, and we will revise accordingly while preserving the core empirical contributions.

read point-by-point responses

  1. Referee: The central claims rest on the assumption that the sampled 4chan and Reddit posts accurately capture primary/secondary resources and expert-to-newcomer knowledge transfer across the broader SNCII ecosystem. No validation, cross-platform checks, or explicit discussion of platform-specific biases (anonymity on 4chan versus moderation on Reddit) or temporal limitations of the 165-day window is provided, leaving open the possibility that observed patterns are artifacts of sampling rather than general features.

    Authors: We acknowledge this observation and will add a dedicated Limitations subsection in the revised manuscript. Our study is explicitly scoped to an integrated analysis of 4chan and Reddit as platforms hosting communities with contrasting user sophistication profiles, as described in the abstract and Section 3. The 165-day collection window (June–November 2025) was selected to capture contemporaneous activity following key platform and regulatory developments; we will now explicitly discuss its temporal bounds and potential for missing longer-term trends. Platform-specific biases (e.g., 4chan’s anonymity versus Reddit’s moderation) are inherent to the data sources and will be addressed by noting how they shape observed resource-sharing behaviors rather than claiming they are absent. Cross-platform validation or external checks are not feasible within ethical and legal constraints on accessing additional underground venues, but we will reference prior smaller-scale studies on SNCII communities to contextualize our large-scale corpus. We do not assert that the patterns generalize to the entire ecosystem; instead, we characterize resource dynamics within these prominent venues. Claims will be tempered and the sampling rationale clarified to address the possibility of artifacts. revision: yes

Circularity Check

0 steps flagged

No circularity: purely observational empirical analysis from forum data

full rationale

The paper performs an integrated analysis of 282,154 4chan comments and 78,308 Reddit submissions over 165 days to characterize actors, actions, and resources in SNCII communities. It reports direct empirical findings on resource sharing and knowledge transfer without any mathematical derivations, fitted parameters, predictions, self-citations, or ansatzes. All claims derive from the collected data itself, with no reduction of outputs to inputs by construction. This is a standard observational study whose central claims rest on the representativeness of the sampled data rather than any self-referential logic.

Axiom & Free-Parameter Ledger

0 free parameters · 2 axioms · 0 invented entities

The central claims rest on assumptions about data representativeness and the interpretation of observed sharing as knowledge transfer; no free parameters or invented entities are introduced.

axioms (2)
  • domain assumption The collected 4chan and Reddit posts represent the primary activities and stakeholders in SNCII creation communities.
    Invoked to generalize from the sampled data to the broader ecosystem.
  • domain assumption Observed patterns of resource mention and user interactions indicate active sharing and knowledge transfer.
    Used to interpret the data as evidence of propagation mechanisms.

pith-pipeline@v0.9.0 · 5546 in / 1410 out tokens · 86799 ms · 2026-05-10T16:12:52.539802+00:00 · methodology

discussion (0)

Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.

Lean theorems connected to this paper

Citations machine-checked in the Pith Canon. Every link opens the source theorem in the public Lean library.

What do these tags mean?
matches
The paper's claim is directly supported by a theorem in the formal canon.
supports
The theorem supports part of the paper's argument, but the paper may add assumptions or extra steps.
extends
The paper goes beyond the formal theorem; the theorem is a base layer rather than the whole result.
uses
The paper appears to rely on the theorem as machinery.
contradicts
The paper's claim conflicts with a theorem or certificate in the canon.
unclear
Pith found a possible connection, but the passage is too broad, indirect, or ambiguous to say the theorem truly supports the claim.

Reference graph

Works this paper leans on

53 extracted references · 53 canonical work pages · 1 internal anchor

  1. [1]

    4chan and/b: An Analysis of Anonymity and Ephemer- ality in a Large Online Community

    Michael Bernstein, Andrés Monroy-Hernández, Drew Harry, Paul André, Katrina Panovich, and Greg Vargas. 4chan and/b: An Analysis of Anonymity and Ephemer- ality in a Large Online Community. InProceedings of the international AAAI conference on web and social media, volume 5, pages 50–57, 2011

    work page 2011

  2. [2]

    Millions created deep- fake nudes on Telegram as AI tools drive global wave of digital abuse

    Priya Bharadia and Aisha Down. Millions created deep- fake nudes on Telegram as AI tools drive global wave of digital abuse. https://www.theguardian.com/gl obal-development/2026/jan/29/millions-cre ating-deepfake-nudes-telegram-ai-digital-a buse, January 2026

    work page 2026

  3. [3]

    Moderating online child sexual abuse material (csam): Does self-regulation work, or is greater state regulation needed?European Journal of Criminol- ogy, 21(2):231–250, 2024

    Paul Bleakley, Elena Martellozzo, Ruth Spence, and Jef- frey DeMarco. Moderating online child sexual abuse material (csam): Does self-regulation work, or is greater state regulation needed?European Journal of Criminol- ogy, 21(2):231–250, 2024

    work page 2024

  4. [4]

    Towards a unified Media-User Typology (MUT): A meta-analysis and review of the research literature on media-user typologies.Computers in Human Behavior, 26(5):940–956, 2010

    Petter Bae Brandtzæg. Towards a unified Media-User Typology (MUT): A meta-analysis and review of the research literature on media-user typologies.Computers in Human Behavior, 26(5):940–956, 2010

    work page 2010

  5. [5]

    Violation of my body:

    Natalie Grace Brigham, Miranda Wei, Tadayoshi Kohno, and Elissa M Redmiles. “Violation of my body:” Per- ceptions of AI-generated non-consensual (intimate) im- agery. InTwentieth Symposium on Usable Privacy and Security (SOUPS 2024), pages 373–392, 2024

    work page 2024

  6. [6]

    SoK: Automated TTP Extraction from CTI Reports–Are We There Yet? In34th USENIX Security Symposium (USENIX Security 25), pages 4621–4641, 2025

    Marvin Büchel, Tommaso Paladini, Stefano Longari, Michele Carminati, Stefano Zanero, Hodaya Binyamini, Gal Engelberg, Dan Klein, Giancarlo Guizzardi, Marco Caselli, et al. SoK: Automated TTP Extraction from CTI Reports–Are We There Yet? In34th USENIX Security Symposium (USENIX Security 25), pages 4621–4641, 2025

    work page 2025

  7. [7]

    Policy update: Removal of real-person likeness content, May 2025

    Civitai. Policy update: Removal of real-person likeness content, May 2025

    work page 2025

  8. [8]

    Porn vilification and age verification: Regulating online pornography and sex work.Feminist Studies, 50(1):40–64, 2024

    Stacey Colliver. Porn vilification and age verification: Regulating online pornography and sex work.Feminist Studies, 50(1):40–64, 2024

    work page 2024

  9. [9]

    Ofcom investigates Elon Musk’s X over Grok AI sexual deepfakes.https: //www.bbc.com/news/articles/cwy875j28k0o , January 2026

    Laura Cress and Liv McMahon. Ofcom investigates Elon Musk’s X over Grok AI sexual deepfakes.https: //www.bbc.com/news/articles/cwy875j28k0o , January 2026

    work page 2026

  10. [10]

    Deep- fake pornography is resilient to regulatory and platform shocks, 2026

    Alejandro Cuevas and Manoel Horta Ribeiro. Deep- fake pornography is resilient to regulatory and platform shocks, 2026. 15

    work page 2026

  11. [11]

    The Data (Use and Access) Act 2025 (Commencement No

    Alex Davies-Jones. The Data (Use and Access) Act 2025 (Commencement No. 5) Regulations 2026.https: //www.legislation.gov.uk/uksi/2026/31/made

    work page 2025

  12. [12]

    Ding and Harini Suresh

    Michelle L. Ding and Harini Suresh. The Malicious Technical Ecosystem: Exposing Limitations in Techni- cal Governance of AI-Generated Non-Consensual Inti- mate Images of Adults, April 2025. arXiv:2504.17663 [cs]

    work page arXiv 2025

  13. [13]

    Remember the Hu- man: A Systematic Review of Ethical Considerations in Reddit Research.Proc

    Casey Fiesler, Michael Zimmer, Nicholas Proferes, Sarah Gilbert, and Naiyan Jones. Remember the Hu- man: A Systematic Review of Ethical Considerations in Reddit Research.Proc. ACM Hum.-Comput. Interact., 8(GROUP):5:1–5:33, February 2024

    work page 2024

  14. [14]

    Understanding learn- ers’ interests in cybersecurity competitions on reddit

    Brock Floyd, Jared Jackson, Emma Probst, Hong Liu, Nischal Mishra, and Chen Zhong. Understanding learn- ers’ interests in cybersecurity competitions on reddit. InProceedings of the 13th International Conference on Education Technology and Computers, pages 444–449, 2021

    work page 2021

  15. [15]

    An Inquiry into the Nature and Causes of the Wealth of Internet Miscreants

    Jason Franklin, Vern Paxson, Adrian Perrig, and Stefan Savage. An Inquiry into the Nature and Causes of the Wealth of Internet Miscreants. InProceedings of the 14th ACM Conference on Computer and Communica- tions Security, CCS ’07, page 375–388, New York, NY , USA, 2007. Association for Computing Machinery

    work page 2007

  16. [16]

    An- alyzing the AI nudification application ecosystem

    Cassidy Gibson, Daniel Olszewski, Natalie Grace Brigham, Anna Crowder, Kevin RB Butler, Patrick Traynor, Elissa M Redmiles, and Tadayoshi Kohno. An- alyzing the AI nudification application ecosystem. In 34th USENIX Security Symposium (USENIX Security 25), pages 1–20, 2025

    work page 2025

  17. [17]

    Deconstructing the Take It Down Act.Commun

    James Grimmelmann. Deconstructing the Take It Down Act.Commun. ACM, 68(9):28–30, August 2025

    work page 2025

  18. [18]

    Characterizing the MrDeepFakes Sexual Deep- fake Marketplace

    Catherine Han, Anne Li, Deepak Kumar, and Zakir Du- rumeric. Characterizing the MrDeepFakes Sexual Deep- fake Marketplace. In34th USENIX Security Symposium (USENIX Security 25), pages 5169–5188, 2025

    work page 2025

  19. [19]

    Deepfakes on Demand: The rise of accessible non- consensual deepfake image generators: The rise of ac- cessible non-consensual deepfake image generators

    Will Hawkins, Brent Mittelstadt, and Chris Russell. Deepfakes on Demand: The rise of accessible non- consensual deepfake image generators: The rise of ac- cessible non-consensual deepfake image generators. In Proceedings of the 2025 ACM Conference on Fairness, Accountability, and Transparency, pages 1602–1614, 2025

    work page 2025

  20. [20]

    ArthurHeitmann/arctic_shift

    Arthur Heitmann. ArthurHeitmann/arctic_shift. https: //github.com/ArthurHeitmann/arctic_shift , January 2026. original-date: 2023-08-03T08:25:15Z

    work page 2026

  21. [21]

    Punishing sex: Sex offenders and the missing punitive turn in sexuality studies.Law & Social Inquiry, 41(3):573–594, 2016

    Trevor Hoppe. Punishing sex: Sex offenders and the missing punitive turn in sexuality studies.Law & Social Inquiry, 41(3):573–594, 2016

    work page 2016

  22. [22]

    Video deepfake abuse: How company choices predictably shape misuse patterns,

    Max Kamachee, Stephen Casper, Michelle L. Ding, Rui-Jie Yew, Anka Reuel, Stella Biderman, and Dylan Hadfield-Menell. Video Deepfake Abuse: How Com- pany Choices Predictably Shape Misuse Patterns, Jan- uary 2026. arXiv:2512.11815 [cs]

    work page arXiv 2026

  23. [23]

    V oelker, Vern Paxson, and Stefan Savage

    Chris Kanich, Christian Kreibich, Kirill Levchenko, Brandon Enright, Geoffrey M. V oelker, Vern Paxson, and Stefan Savage. Spamalytics: an empirical analy- sis of spam marketing conversion. InProceedings of the 15th ACM Conference on Computer and Communi- cations Security, CCS ’08, page 3–14, New York, NY , USA, 2008. Association for Computing Machinery

    work page 2008

  24. [24]

    On assess- ing interrater agreement for multiple attribute responses

    Lawrence L Kupper and Kerry B Hafner. On assess- ing interrater agreement for multiple attribute responses. Biometrics, pages 957–967, 1989

    work page 1989

  25. [25]

    PharmaLeaks: Understanding the Business of Online Pharmaceutical Affiliate Programs

    Damon McCoy, Andreas Pitsillidis, Jordan Grant, Nicholas Weaver, Christian Kreibich, Brian Krebs, Ge- offrey V oelker, Stefan Savage, and Kirill Levchenko. PharmaLeaks: Understanding the Business of Online Pharmaceutical Affiliate Programs. In21st USENIX Security Symposium (USENIX Security 12), pages 1–16, Bellevue, W A, August 2012. USENIX Association

    work page 2012

  26. [26]

    Can deepfakes be created on a whim? InCompanion Proceedings of the ACM Web Conference 2023, WWW ’23 Companion, page 1324–1334, New York, NY , USA, 2023

    Pulak Mehta, Gauri Jagatap, Kevin Gallagher, Brian Timmerman, Progga Deb, Siddharth Garg, Rachel Greenstadt, and Brendan Dolan-Gavitt. Can deepfakes be created on a whim? InCompanion Proceedings of the ACM Web Conference 2023, WWW ’23 Companion, page 1324–1334, New York, NY , USA, 2023. Associa- tion for Computing Machinery

    work page 2023

  27. [27]

    Unlim- ited Realm of Exploration and Experimentation

    Jaron Mink, Lucy Qin, and Elissa M Redmiles. “Unlim- ited Realm of Exploration and Experimentation”: Meth- ods and Motivations of AI-Generated Sexual Content Creators.arXiv preprint arXiv:2601.21028, 2026

    work page internal anchor Pith review arXiv 2026

  28. [28]

    V oelker, and Kurt Thomas

    Ariana Mirian, Joe DeBlasio, Stefan Savage, Geof- frey M. V oelker, and Kurt Thomas. Hack for Hire: Ex- ploring the Emerging Market for Account Hijacking. InThe World Wide Web Conference, WWW ’19, page 1279–1289, New York, NY , USA, 2019. Association for Computing Machinery

    work page 2019

  29. [29]

    Taylor Swift AI images prompt US bill to tackle nonconsensual, sexual deepfakes

    Blake Montgomery. Taylor Swift AI images prompt US bill to tackle nonconsensual, sexual deepfakes. https: //www.theguardian.com/technology/2024/jan/ 30/taylor-swift-ai-deepfake-nonconsensual -sexual-images-bill, January 2024. 16

    work page 2024

  30. [30]

    Re:{CAPTCHAs—Understanding }{CAPTCHA- Solving} services in an economic context

    Marti Motoyama, Kirill Levchenko, Chris Kanich, Da- mon McCoy, Geoffrey M V oelker, and Stefan Sav- age. Re:{CAPTCHAs—Understanding }{CAPTCHA- Solving} services in an economic context. In19th USENIX Security Symposium (USENIX Security 10), 2010

    work page 2010

  31. [31]

    V oelker

    Marti Motoyama, Damon McCoy, Kirill Levchenko, Ste- fan Savage, and Geoffrey M. V oelker. An analysis of underground forums. InProceedings of the 2011 ACM SIGCOMM Conference on Internet Measurement Con- ference, IMC ’11, page 71–80, New York, NY , USA,

    work page 2011

  32. [32]

    Association for Computing Machinery

    work page

  33. [33]

    Apple, Google host dozens of AI ‘nudify’ apps like Grok, report finds

    Lola Murti. Apple, Google host dozens of AI ‘nudify’ apps like Grok, report finds. https://www.cnbc.com /2026/01/27/apple-google-host-dozens-of-a i-nudify-apps-like-grok-report-finds.html , January 2026. Section: Technology

    work page 2026

  34. [34]

    Victims, Vigilantes, and Advice Givers: An Analysis of {Scam-Related} Discourse on Reddit

    Rajvardhan Oak and Zubair Shafiq. Victims, Vigilantes, and Advice Givers: An Analysis of {Scam-Related} Discourse on Reddit. InTwenty-First Symposium on Usable Privacy and Security (SOUPS 2025), pages 57– 71, 2025

    work page 2025

  35. [35]

    Automatic mapping of unstructured cyber threat intelligence: An experimen- tal study:(practical experience report)

    Vittorio Orbinato, Mariarosaria Barbaraci, Roberto Natella, and Domenico Cotroneo. Automatic mapping of unstructured cyber threat intelligence: An experimen- tal study:(practical experience report). In2022 IEEE 33rd International Symposium on Software Reliability Engineering (ISSRE), pages 181–192. IEEE, 2022

    work page 2022

  36. [36]

    Lei nº 15.123, de 2025

    Presidência da República Federativa do Brasil. Lei nº 15.123, de 2025. https://www.planalto.gov.br/cc ivil_03/_ato2023-2026/2025/Lei/L15123.htm ,

    work page 2025

  37. [37]

    Accessed: 2026-01-22

    work page 2026

  38. [38]

    The ecosystem of nonconsensual intimate deepfake tools online

    Chiara Puglielli and Anne Craanen. The ecosystem of nonconsensual intimate deepfake tools online. https: //www.isdglobal.org/digital_dispatches/the -ecosystem-of-nonconsensual-intimate-deepf ake-tools-online/, October 2025

    work page 2025

  39. [39]

    did they {F*** ing} consent to that?

    Lucy Qin, Vaughn Hamilton, Sharon Wang, Yigit Ay- dinalp, Marin Scarlett, and Elissa M Redmiles. " did they {F*** ing} consent to that?": Safer digital intimacy via proactive protection against {Image-Based} sexual abuse. In33rd USENIX Security Symposium (USENIX Security 24), pages 55–72, 2024

    work page 2024

  40. [40]

    Beautiful soup documentation

    Leonard Richardson. Beautiful soup documentation. https://www.crummy.com/software/BeautifulS oup/bs4/doc/, 2007

    work page 2007

  41. [41]

    Classifying cyber-risky clinical notes by employing natural language processing

    Suzanna Schmeelk, Martins Samuel Dogo, Yifan Peng, and Braja Gopal Patra. Classifying cyber-risky clinical notes by employing natural language processing. InPro- ceedings of the Annual Hawaii International Conference on System Sciences, page 4140, 2022

    work page 2022

  42. [42]

    Ted [R-TX Sen. Cruz. S.146 - 119th Congress (2025- 2026): TAKE IT DOWN Act. https://www.congre ss.gov/bill/119th-congress/senate-bill/146 , May 2025. Archive Location: 2025-01-16

    work page 2025

  43. [43]

    Mandatory age verification for pornography access: Why it can’t and won’t ‘save the children’.Big data & society, 11(2):20539517241252129, 2024

    Zahra Stardust, Abdul Obeid, Alan McKee, and Daniel Angus. Mandatory age verification for pornography access: Why it can’t and won’t ‘save the children’.Big data & society, 11(2):20539517241252129, 2024

    work page 2024

  44. [44]

    Yuan Y . Stevens. Canada Must Do More to Protect Women and Girls from Harmful Deepfakes. https: //www.cigionline.org/articles/canada-mus t-do-more-to-protect-women-and-girls-fro m-harmful-deepfakes/

    work page

  45. [45]

    Your botnet is my botnet: analysis of a botnet takeover

    Brett Stone-Gross, Marco Cova, Lorenzo Cavallaro, Bob Gilbert, Martin Szydlowski, Richard Kemmerer, Christo- pher Kruegel, and Giovanni Vigna. Your botnet is my botnet: analysis of a botnet takeover. InProceedings of the 16th ACM Conference on Computer and Commu- nications Security, CCS ’09, page 635–647, New York, NY , USA, 2009. Association for Computin...

    work page 2009

  46. [46]

    Beverly Hills middle school expels 5 students after deepfake nude photos incident

    Kat Tenbarge. Beverly Hills middle school expels 5 students after deepfake nude photos incident. https: //www.nbcnews.com/tech/tech-news/beverly-h ills-school-expels-students-deepfake-nud e-photos-rcna142480, March 2024

    work page 2024

  47. [47]

    Think Twice Before You Search: Deterrence Messaging Designs to Prevent Searches for Non-Consensual Intimate Images

    Dhanaraj Thakur and Varun Nagaraj Rao. Think Twice Before You Search: Deterrence Messaging Designs to Prevent Searches for Non-Consensual Intimate Images. https://cdt.org/insights/think-twice-befor e-you-search-deterrence-messaging-designs -to-prevent-searches-for-non-consensual-i ntimate-images, November 2025

    work page 2025

  48. [48]

    Studying the online deepfake com- munity.Journal of Online Trust and Safety, 2(1), 2023

    Brian Timmerman, Pulak Mehta, Progga Deb, Kevin Gallagher, Brendan Dolan-Gavitt, Siddharth Garg, and Rachel Greenstadt. Studying the online deepfake com- munity.Journal of Online Trust and Safety, 2(1), 2023

    work page 2023

  49. [49]

    arXiv preprint arXiv:2209.11055 , year=

    Lewis Tunstall, Nils Reimers, Unso Eun Seo Jo, Luke Bates, Daniel Korat, Moshe Wasserblat, and Oren Pereg. Efficient few-shot learning without prompts.arXiv preprint arXiv:2209.11055, 2022

    work page arXiv 2022

  50. [50]

    Non-consensual synthetic intimate imagery: Prevalence, attitudes, and knowledge in 10 countries

    Rebecca Umbach, Nicola Henry, Gemma Faye Beard, and Colleen M Berryessa. Non-consensual synthetic intimate imagery: Prevalence, attitudes, and knowledge in 10 countries. InProceedings of the 2024 CHI Confer- ence on Human Factors in Computing Systems, pages 1–20, 2024. 17

    work page 2024

  51. [51]

    Understanding Help-Seeking and Help-Giving on social media for Image-Based sexual abuse

    Miranda Wei, Sunny Consolvo, Patrick Gage Kelley, Tadayoshi Kohno, Tara Matthews, Sarah Meiklejohn, Franziska Roesner, Renee Shelby, Kurt Thomas, and Rebecca Umbach. Understanding Help-Seeking and Help-Giving on social media for Image-Based sexual abuse. In33rd USENIX Security Symposium (USENIX Security 24), pages 4391–4408, 2024

    work page 2024

  52. [52]

    We’re utterly ill-prepared to deal with something like this

    Miranda Wei, Christina Yeung, Franziska Roesner, and Tadayoshi Kohno. “We’re utterly ill-prepared to deal with something like this”: Teachers’ perspectives on student generation of synthetic nonconsensual explicit imagery. InProceedings of the 2025 CHI Conference on Human Factors in Computing Systems, 2025

    work page 2025

  53. [53]

    r=”, “ref=

    Qunfang Wu, Louisa Kayah Williams, Ellen Simpson, and Bryan Semaan. Conversations about crime: Re- enforcing and fighting against platformed racism on reddit.Proceedings of the ACM on Human-Computer Interaction, 6(CSCW1):1–38, 2022. Appendix A Methodology We provide here further detail on our data collection and analysis methodologies for reproducibility ...

    work page 2022