CoChat
Sign Up
Cited by 11Ethics and Social Impacts of AIHate Speech and Cyberbullying DetectionMachine Learning and Data Classification

Who's in and who's out? A case study of multimodal CLIP-filtering in DataComp

Open access
Hong, Rachel, Agnew, William, Kohno, Tadayoshi, Morgenstern, Jamie·2024·Proceedings of the 4th ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization
DOI 10.1145/3689904.3694702
Open this paper in CoChat — free account

Abstract

As training datasets become increasingly drawn from unstructured, uncontrolled environments such as the web, researchers and industry practitioners have increasingly relied upon data filtering techniques to “filter out the noise” of web-scraped data. While datasets have been widely shown to reflect the biases and values of their creators, in this paper we contribute to an emerging body of research that assesses the filters used to create these datasets. We show that image-text data filtering also has biases and is value-laden, encoding specific notions of what is counted as “high-quality” data. In our work, we audit a standard approach of image-text CLIP-filtering on the academic benchmark DataComp’s CommonPool by analyzing discrepancies of filtering through various annotation techniques across multiple modalities of image, text, and website source. We find that data relating to several imputed demographic groups — such as LGBTQ+ people, older women, and younger men — are associated with higher rates of exclusion. We also find prevalence of Western bias, where the CLIP filter is more likely to include data related to Western countries compared to that of non-Western countries. Moreover, we demonstrate cases of exclusion amplification: not only are certain marginalized groups already underrepresented in the unfiltered data, but CLIP-filtering excludes data from these groups at higher rates. The data-filtering step in the machine learning pipeline can therefore exacerbate representation disparities already present in the data-gathering step, especially when existing filters are designed to optimize a specifically-chosen downstream performance metric like zero-shot image classification accuracy. Finally, we show that the NSFW filter fails to remove sexually-explicit content from CommonPool, and that CLIP-filtering includes several categories of copyrighted content at high rates. Our conclusions point to a need for fundamental changes in dataset creation and filtering practices. Content warning: This paper discusses societal stereotypes and sexually-explicit material that may be disturbing, distressing, and/or offensive to the reader.

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 · 10.1145/3689904.3694702

Cites

15 of 15 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?

  • Current data collection processes prioritize scale over inclusion, relying on web-scraping and synthetic pipelines that reflect developer biases and the uneven availability of online content.

    “Current data collection processes prioritize scale over inclusion, relying on web-scraping (e.g., [40, 97]) and synthetic pipelines that reflect developer biases and the uneven availability of online content [13, 14, 57, 96].”

    From Engaging Communities Meaningfully in Defining Disability Representation for AI Image Generation (2026) · cites this paper as [57]

    The Introduction (p. 1) says datasets are scraped from the web to reach a scale of billions, and the Abstract (p. 1) reports bias and amplified underrepresentation in the resulting filtered data. But this paper studies filtering of web-scraped data, not synthetic data pipelines; it does not support that part of the claim. (LLM, conf=0.87)

    Experimental — accuracy not yet measured

  • Even when people with disabilities data is available, data from marginalized groups is often automatically filtered out by techniques for removing 'low' quality data, which rely on biased models such as CLIP.

    “Even when PwD data is available, data from marginalized groups is often automatically filtered out by techniques for removing 'low' quality data [57, 104], which rely on biased models such as CLIP [92].”

    From Engaging Communities Meaningfully in Defining Disability Representation for AI Image Generation (2026) · cites this paper as [57]

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: 69

  1. [1][1] Sandhini Agarwal, Gretchen Krueger, Jack Clark, Alec Radford, Jong Wook Kim, and Miles Brundage. 2021. Evaluating CLIP: Towards Characterization of Broader Capabilities and Downstream Implications. http://arxiv.org/abs/2108. 02818 arXiv:2108.02818 [cs]. - · arXiv
  2. [3][3] Jacob Alhadeff, Cooper Cuene, and Max Del Real. 2024. Limits of Algorithmic Fair Use. Wash. JL Tech. & Arts 19 (2024), 1. -
  3. [4][4] Junaid Ali, Matthäus Kleindessner, Florian Wenzel, Kailash Budhathoki, Volkan Cevher, and Chris Russell. 2023. Evaluating the Fairness of Discriminative Foundation Models in Computer Vision. In Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society . ACM, Montreal, QC, Canada, 809-833. https://doi.org/10.1145/3600211.3604720 - · DOI
  4. [6][6] Jerone Andrews, Dora Zhao, William Thong, Apostolos Modas, Orestis Papakyriakopoulos, and Alice Xiang. 2024. Ethical Considerations for Responsible Data Curation. Advances in Neural Information Processing Systems 36 (2024), 55320-55360. -
  5. [8][8] Sameena Azhar, Antonia RG Alvarez, Anne SJ Farina, and Susan Klumpner. 2021. 'You're so exotic looking': An intersectional analysis of Asian American and Pacific Islander stereotypes. Affilia 36, 3 (2021), 282-301. -
  6. [9][9] Andy Baio. 2022. Exploring 12 million of the 2.3 billion images used to train stable diffusion's image generator. Retrieved July 6 (2022), 2023. https://waxy.org/2022/08/exploring-12-million-of-the-images-used-totrain-stable-diffusions-image-generator/ -
  7. [11][11] Federico Bianchi, Pratyusha Kalluri, Esin Durmus, Faisal Ladhak, Myra Cheng, Debora Nozza, Tatsunori Hashimoto, Dan Jurafsky, James Zou, and Aylin Caliskan. 2023. Easily Accessible Text-to-Image Generation Amplifies Demographic Stereotypes at Large Scale. In 2023 ACM Conference on Fairness, Accountability, and Transparency . ACM, Chicago, IL, USA, 1493-1504. https: //doi.org/10.1145/3593013.3594095 - · DOI · Paper page
  8. [13][13] Charlotte Bird, Eddie Ungless, and Atoosa Kasirzadeh. 2023. Typology of risks of generative text-to-image models. In Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society . ACM, Montreal, Canada, 396-410. -
  9. [14][14] Abeba Birhane, Pratyusha Kalluri, Dallas Card, William Agnew, Ravit Dotan, and Michelle Bao. 2022. The values encoded in machine learning research. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency . ACM, Seoul, South Korea, 173-184. -
  10. [16][16] Abeba Birhane and Vinay Uday Prabhu. 2021. Large image datasets: A pyrrhic win for computer vision?. In 2021 IEEE Winter Conference on Applications of Computer Vision (WACV) . IEEE, Online, 1536-1546. -
  11. [18][18] SumonBiswas and Hridesh Rajan. 2021. Fair preprocessing: towards understanding compositional fairness of data transformers in machine learning pipeline. In Proceedings of the 29th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering . ACM, Athens, Greece, 981-993. https://doi.org/10.1145/3468264.3468536 - · DOI
  12. [21][21] Brandon M Booth, Louis Hickman, Shree Krishna Subburaj, Louis Tay, Sang Eun Woo, and Sidney K D'Mello. 2021. Bias and fairness in multimodal machine learning: A case study of automated video interviews. In Proceedings of the 2021 International Conference on Multimodal Interaction . ACM, Montreal, QC, Canada, 268-277. -
  13. [23][23] Dawn Beverley Branley and Judith Covey. 2017. Is exposure to online content depicting risky behavior related to viewers' own risky behavior offline? Computers in Human Behavior 75 (2017), 283-287. -
  14. [24][24] Tom Brown, Benjamin Mann, Nick Ryder, Melanie Subbiah, Jared D Kaplan, Prafulla Dhariwal, Arvind Neelakantan, Pranav Shyam, Girish Sastry, Amanda Askell, et al. 2020. Language models are few-shot learners. Advances in neural information processing systems 33 (2020), 1877-1901. -
  15. [25][25] Joy Buolamwini and Timnit Gebru. 2018. Gender shades: Intersectional accuracy disparities in commercial gender classification. In Conference on Fairness, Accountability, and Transparency . PMLR, New York, NY, USA, 77-91. -
  16. [26][26] Judith Butler. 2013. Gender as performance. In A critical sense . Routledge, New York, NY, USA, 109-125. -
  17. [28][28] Ben Caldwell, Michael Cooper, Loretta Guarino Reid, Gregg Vanderheiden, Wendy Chisholm, John Slatin, and Jason White. 2008. Web content accessibility guidelines (WCAG) 2.0. WWWConsortium (W3C) 290 (2008), 1-34. -
  18. [29][29] Aylin Caliskan, Pimparkar Parth Ajay, Tessa Charlesworth, Robert Wolfe, and Mahzarin R Banaji. 2022. Gender bias in word embeddings: A comprehensive analysis of frequency, syntax, and semantics. In Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society . ACM, Oxford, England, 156170. -
  19. [30][30] Irene Chen, Fredrik D Johansson, and David Sontag. 2018. Why is my classifier discriminatory? Advances in neural information processing systems 31 (2018). -
  20. [38][38] Katherine Crowson, Stella Biderman, Daniel Kornis, Dashiell Stander, Eric Hallahan, Louis Castricato, and Edward Raff. 2022. Vqgan-clip: Open domain image generation and editing with natural language guidance. In European Conference on Computer Vision . Springer, Tel Aviv, Israel, 88-105. -
  21. [39][39] Allan Dafoe. 2015. On technological determinism: A typology, scope conditions, and a mechanism. Science, Technology, & Human Values 40, 6 (2015), 1047-1076. -
  22. [45][45] Cynthia Dwork, Moritz Hardt, Toniann Pitassi, Omer Reingold, and Richard Zemel. 2012. Fairness through awareness. In Proceedings of the 3rd Innovations in Theoretical Computer Science Conference . ACM, Cambridge, MA, USA, 214-226. -
  23. [48][48] Hany Farid. 2021. An overview of perceptual hashing. Journal of Online Trust and Safety 1, 1 (2021), 22 pages. -
  24. [50][50] Sachin Goyal, Pratyush Maini, Zachary C. Lipton, Aditi Raghunathan, and J. Zico Kolter. 2024. Scaling Laws for Data Filtering - Data Curation cannot be Compute Agnostic. arXiv:2404.07177 [cs.LG] - · arXiv
  25. [52][52] Shubha Guha, Falaah Arif Khan, Julia Stoyanovich, and Sebastian Schelter. 2023. Automated Data Cleaning Can Hurt Fairness in Machine Learningbased Decision Making. In 2023 IEEE 39th International Conference on Data Engineering (ICDE) . IEEE, Anaheim, CA, USA, 3747-3754. https://doi.org/10. 1109/ICDE55515.2023.00303 -
  26. [54][54] Ritwik Gupta. 2024. LAION and the Challenges of Preventing AI-Generated CSAM. https://www.techpolicy.press/laion-and-the-challenges-of-preventingai-generated-csam/ -
  27. [57][57] Peter Henderson, Xuechen Li, Dan Jurafsky, Tatsunori Hashimoto, Mark A. Lemley, and Percy Liang. 2023. Foundation Models and Fair Use. https: //doi.org/10.48550/arXiv.2303.15715 arXiv:2303.15715 [cs]. - · DOI
  28. [58][58] Mariya Hendriksen, Maurits Bleeker, Svitlana Vakulenko, Nanne van Noord, Ernst Kuiper, and Maarten de Rijke. 2022. Extending CLIP for Category-toimage Retrieval in E-commerce. In European Conference on Information Retrieval . Springer, Stavanger, Norway, 289-303. -
  29. [60][60] Ben Hutchinson, Andrew Smart, Alex Hanna, Emily Denton, Christina Greer, Oddur Kjartansson, Parker Barnes, and Margaret Mitchell. 2021. Towards accountability for machine learning datasets: Practices from software engineering and infrastructure. In Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency . ACM, Online, 560-575. -
  30. [63][63] Eun Seo Jo and Timnit Gebru. 2020. Lessons from archives: Strategies for collecting sociocultural data in machine learning. In Proceedings of the 2020 conference on fairness, accountability, and transparency . ACM, Barcelona, Spain, 306-316. -
  31. [65][65] Mehtab Khan and Alex Hanna. 2022. The subjects and stages of ai dataset development: A framework for dataset accountability. Ohio St. Tech. LJ 19 (2022), 171. -
  32. [66][66] Dan Komosny, Miroslav Voznak, and Saeed Ur Rehman. 2017. Location accuracy of commercial IP address geolocation databases. Information technology and control 46, 3 (2017), 333-344. -
  33. [67][67] Matt J Kusner, Joshua Loftus, Chris Russell, and Ricardo Silva. 2017. Counterfactual fairness. Advances in neural information processing systems 30 (2017), 11 pages. -
  34. [68][68] Ganaele Langlois and Andrea Slane. 2017. Economies of reputation: The case of revenge porn. Communication and Critical/Cultural Studies 14, 2 (2017), 120-138. -
  35. [69][69] Hee-Eun Lee, Tatiana Ermakova, Vasilis Ververis, and Benjamin Fabian. 2020. Detecting child sexual abuse material: A comprehensive survey. Forensic Science International: Digital Investigation 34 (2020), 301022. -
  36. [70][70] Katherine Lee, A. Feder Cooper, and James Grimmelmann. 2023. Talkin' 'Bout AI Generation: Copyright and the Generative-AI Supply Chain. https://doi. org/10.2139/ssrn.4523551 - · DOI
  37. [71][71] Jonathan D Levine and Frederick L Oswald. 2013. O* NET: The occupational information network. In The Handbook of Work Analysis . Routledge, New York, NY, USA, 312-332. -
  38. [72][72] Ioana Livadariu, Thomas Dreibholz, Anas Saeed Al-Selwi, Haakon Bryhni, Olav Lysne, Steinar Bjørnstad, and Ahmed Elmokashfi. 2020. On the accuracy of country-level IP geolocation. In Proceedings of the applied networking research workshop . ACM, Online, 67-73. -
  39. [73][73] Ian F Haney Lopez. 1995. The social construction of race . Harvard Civil RightsCivil Liberties Law Review, Cambridge, MA, USA. -
  40. [74][74] Alexandra Sasha Luccioni, Christopher Akiki, Margaret Mitchell, and Yacine Jernite. 2024. Stable bias: Evaluating societal representations in diffusion models. Advances in Neural Information Processing Systems 36 (2024), 56338-56351. -
  41. [75][75] Alexandra Sasha Luccioni, Frances Corry, Hamsini Sridharan, Mike Ananny, Jason Schultz, and Kate Crawford. 2022. A framework for deprecating datasets: Standardizing documentation, identification, and communication. In Proceedings of the 2022 ACM Conference on Fairness, Accountability, and Transparency . ACM, Seoul, South Korea, 199-212. -
  42. [76][76] Alexandra Sasha Luccioni and Joseph Viviano. 2021. What's in the box? An analysis of undesirable content in the Common Crawl corpus. In Proceedings of the 59th Annual Meeting of the Association for Computational Linguistics and the 11th International Joint Conference on Natural Language Processing (Volume 2: Short Papers) . ACL, Bangkok, Thailand, 182-189. -
  43. [78][78] Debbie S Ma, Joshua Correll, and Bernd Wittenbrink. 2015. The Chicago face database: A free stimulus set of faces and norming data. Behavior research methods 47 (2015), 1122-1135. -
  44. [85][85] Arif Ali Mughal. 2018. The Art of Cybersecurity: Defense in Depth Strategy for Robust Protection. International Journal of Intelligent Automation and Computing 1, 1 (2018), 1-20. -
  45. [86][86] Michael Muller and Angelika Strohmayer. 2022. Forgetting practices in the data sciences. In Proceedings of the 2022 CHI Conference on Human Factors in Computing Systems . ACM, New Orleans, LA, USA, 1-19. -
  46. [88][88] Thao Nguyen, Gabriel Ilharco, Mitchell Wortsman, Sewoong Oh, and Ludwig Schmidt. 2022. Quality not quantity: On the interaction between dataset design and robustness of clip. Advances in Neural Information Processing Systems 35 (2022), 21455-21469. -
  47. [91][91] Maya Okawa, Ekdeep S Lubana, Robert Dick, and Hidenori Tanaka. 2024. Compositional abilities emerge multiplicatively: Exploring diffusion models on a synthetic task. Advances in Neural Information Processing Systems 36 (2024), 28. 23 pages. -
  48. [92][92] Ruth Oldenziel. 1999. Making technology masculine: men, women and modern machines in America, 1870-1945 . Amsterdam University Press, Amsterdam, Netherlands. -
  49. [96][96] Ingmar Poese, Steve Uhlig, Mohamed Ali Kaafar, Benoit Donnet, and Bamba Gueye. 2011. IP geolocation databases: Unreliable? ACM SIGCOMM Computer Communication Review 41, 2 (2011), 53-56. -
  50. [97][97] Haoyi Qiu, Zi-Yi Dou, Tianlu Wang, Asli Celikyilmaz, and Nanyun Peng. 2023. Gender Biases in Automatic Evaluation Metrics for Image Captioning. In Proceedings of the 2023 Conference on Empirical Methods in Natural Language Processing , Houda Bouamor, Juan Pino, and Kalika Bali (Eds.). Association for Computational Linguistics, Singapore, 8358-8375. https://doi.org/10.18653/v1/ 2023.emnlp-main.520 - · DOI
  51. [98][98] Yiting Qu, Xinyue Shen, Xinlei He, Michael Backes, Savvas Zannettou, and Yang Zhang. 2023. Unsafe diffusion: On the generation of unsafe images and hateful memes from text-to-image models. In Proceedings of the 2023 ACM SIGSAC Conference on Computer and Communications Security . ACM, Copenhagen, Denmark, 3403-3417. -
  52. [99][99] Alec Radford, Jong Wook Kim, Chris Hallacy, Aditya Ramesh, Gabriel Goh, Sandhini Agarwal, Girish Sastry, Amanda Askell, Pamela Mishkin, Jack Clark, et al. 2021. Learning transferable visual models from natural language supervision. In International Conference on Machine Learning . PMLR, Online, 8748-8763. -
  53. [100][100] Alec Radford, Jeffrey Wu, Rewon Child, David Luan, Dario Amodei, Ilya Sutskever, et al. 2019. Language models are unsupervised multitask learners. OpenAI blog 1, 8 (2019), 9. -
  54. [101][101] Colin Raffel, Noam Shazeer, Adam Roberts, Katherine Lee, Sharan Narang, Michael Matena, Yanqi Zhou, Wei Li, and Peter J Liu. 2020. Exploring the limits of transfer learning with a unified text-to-text transformer. The Journal of Machine Learning Research 21, 1 (2020), 5485-5551. -
  55. [102][102] World Population Review. 2024. Western Countries 2024. worldpopulationreview.com/country-rankings/western-countries [https://](https://worldpopulationreview.com/country-rankings/western-countries) -
  56. [104][104] Esther Rolf, Theodora T Worledge, Benjamin Recht, and Michael Jordan. 2021. Representation matters: Assessing the importance of subgroup allocations in training data. In International Conference on Machine Learning . PMLR, Online, 9040-9051. -
  57. [106][106] Kimberly Ruth, Aurore Fass, Jonathan Azose, Mark Pearson, Emma Thomas, Caitlin Sadowski, and Zakir Durumeric. 2022. A world wide view of browsing the world wide web. In Proceedings of the 22nd ACM Internet Measurement Conference . ACM, Nice, France, 317-336. -
  58. [107][107] Aneeshan Sain, Ayan Kumar Bhunia, Pinaki Nath Chowdhury, Subhadeep Koley, Tao Xiang, and Yi-Zhe Song. 2023. Clip for all things zero-shot sketch-based image retrieval, fine-grained or not. In Proceedings of the IEEE/CVF Conference on Computer Vision and Pattern Recognition . ACM, Vancouver, Canada, 2765-2775. -
  59. [110][110] Morgan Klaus Scheuerman, Alex Hanna, and Emily Denton. 2021. Do datasets have politics? Disciplinary values in computer vision dataset development. Proceedings of the ACM on Human-Computer Interaction 5, CSCW2 (2021), 1-37. -
  60. [111][111] Christoph Schuhmann, Romain Beaumont, Richard Vencu, Cade Gordon, Ross Wightman, Mehdi Cherti, Theo Coombes, Aarush Katta, Clayton Mullis, Mitchell Wortsman, et al. 2022. Laion-5b: An open large-scale dataset for training next generation image-text models. Advances in Neural Information Processing Systems 35 (2022), 25278-25294. -
  61. [113][113] Carsten Schwemmer, Carly Knight, Emily D Bello-Pardo, Stan Oklobdzija, Martijn Schoonvelde, and Jeffrey W Lockhart. 2020. Diagnosing gender bias in image recognition systems. Socius 6 (2020), 2378023120967171. -
  62. [114][114] Renee Shelby, Shalaleh Rismani, Kathryn Henne, AJung Moon, Negar Rostamzadeh, Paul Nicholas, N'Mah Yilla-Akbari, Jess Gallegos, Andrew Smart, Emilio Garcia, et al. 2023. Sociotechnical harms of algorithmic systems: Scoping - a taxonomy for harm reduction. In Proceedings of the 2023 AAAI/ACM Conference on AI, Ethics, and Society . ACM, Montreal, QC, Canada, 723-741. -
  63. [118][118] Morgan P Slusher and Craig A Anderson. 1987. When reality monitoring fails: The role of imagination in stereotype maintenance. Journal of Personality and Social Psychology 52, 4 (1987), 653. -
  64. [124][124] Yonglong Tian, Dilip Krishnan, and Phillip Isola. 2020. Contrastive multiview coding. In European Conference on Computer Vision . Springer, Online, 776-794. -
  65. [125][125] Francisco Valdes. 1996. Unpacking hetero-patriarchy: tracing the conflation of sex, gender & (and) sexual orientation to its origins. Yale JL & Human. 8 (1996), 161. -
  66. [131][131] Kaylee Williams. 2023. Exploring Legal Approaches to Regulating Nonconsensual Deepfake Pornography. https://www.techpolicy.press/exploring-legalapproaches-to-regulating-nonconsensual-deepfake-pornography/ -
  67. [133][133] Robert Wolfe and Aylin Caliskan. 2022. American == White in Multimodal Language-and-Image AI. In Proceedings of the 2022 AAAI/ACM Conference on AI, Ethics, and Society . ACM, Oxford, United Kingdom, 800-812. https://doi. org/10.1145/3514094.3534136 - · DOI
  68. [134][134] Robert Wolfe, Yiwei Yang, Bill Howe, and Aylin Caliskan. 2023. Contrastive language-vision ai models pretrained on web-scraped multimodal data exhibit sexual objectification bias. In Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency . ACM, Chicago, IL, USA, 1174-1185. -
  69. [137][137] Jiping Zuo and Shengming Tang. 2000. Breadwinner status and gender ideologies of men and women regarding family roles. Sociological perspectives 43, 1 (2000), 29-43.

Cited by

Sorted by citations

Authors
Hong, R., Agnew, W., Kohno, T., & Morgenstern, J.
Affiliations
  • University of Washington (US)
  • Carnegie Mellon University (US)
ORCID
Venue
Proceedings of the 4th ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization
Publisher
ACM
Published
2024-10-29
Pages
1-17
DOI
10.1145/3689904.3694702
OpenAlex
W4403681535
Cited by
11
References
15
Citations by year
2026: 6 · 2025: 5
FWCI
5.01
Type
conference paper
Language
en
Open access
Licence
Locations
  • ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization (EAAMO) (publishedVersion, cc-by)
  • arXiv (Cornell University) (submittedVersion, cc-by)
  • arXiv (Cornell University)
Funders
National Science Foundation (CNS-2205171, CCF-2045402, GRFP)
Topics
  • Social Sciences › Safety Research › Ethics and Social Impacts of AI
  • Physical Sciences › Computer Science › Artificial Intelligence › Hate Speech and Cyberbullying Detection
  • Physical Sciences › Computer Science › Artificial Intelligence › Machine Learning and Data Classification
Keywords
data filtering, image-text dataset, data bias, demographic bias, Western bias

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