Prévalence des biais à l'encontre des termes liés à la neurodivergence dans les modèles de langage d'intelligence artificielle

Titre original en anglais : Prevalence of bias against neurodivergence‐related terms in artificial intelligence language models

Sources de métadonnées utilisées pour cette référence

OpenAlex Crossref OpenCitations
En savoir plus sur les sources utilisées

Brandsen, S., Chandrasekhar, T., Franz, L., Grapel, J., Dawson, G., & Carlson, D. (2024). Prevalence of bias against neurodivergence‐related terms in artificial intelligence language models. Autism Research, 17(2), 234-248. https://doi.org/10.1002/aur.3094

Date de publication: 29/01/2024 Ajout dans AutiHub: 05/07/2026 Type: Article Langue de l’article: Anglais
Source OpenAlex Source éditeur (DOI) Texte intégral

Cette publication est intégrée dans AutiHub via :

Auteurs

Auteur·ices des publications
6
Auteur·ices de la publication identifié·es comme autistes
2 / 6 (33,3 %)
  1. Sam Brandsen *
  2. Tara Chandrasekhar
  3. Lauren Franz
  1. Jordan Grapel *
  2. Géraldine Dawson
  3. David Carlson

Résumé

Compte tenu du rôle croissant de l'intelligence artificielle (IA) dans de nombreux processus de prise de décision, nous étudions la présence de biais de l'IA envers des termes liés à un éventail de conditions neurodivergentes, notamment l'autisme, le TDAH, la schizophrénie et le trouble obsessionnel-compulsif (TOC). Nous utilisons 11 encodeurs de modèles de langage différents pour tester le degré auquel des mots liés à la neurodiversité sont associés à des groupes de mots liés au danger, à la maladie, à la méchanceté et à d'autres concepts négatifs. Pour chaque groupe de mots testé, nous rapportons la force moyenne d'association (score du Word Embedding Association Test [WEAT]) calculée sur l'ensemble des encodeurs et constatons des niveaux de biais généralement élevés. De plus, nous montrons que les biais se produisent même lors du test de mots associés aux forces autistiques ou neurodivergentes. Par exemple, les encodeurs présentaient une association moyenne négative entre des mots liés à l'autisme et des mots liés à l'honnêteté, malgré le fait que l'honnêteté soit considérée comme une force courante des personnes autistes. Enfin, nous introduisons un test de ratio de similarité des phrases et démontrons que de nombreuses phrases décrivant des types de handicaps, par exemple, « J'ai de l'autisme » ou « J'ai de l'épilepsie », présentent des associations négatives encore plus fortes que des phrases témoins telles que « Je suis un braqueur de banque ».

Given the increasing role of artificial intelligence (AI) in many decision-making processes, we investigate the presence of AI bias towards terms related to a range of neurodivergent conditions, including autism, ADHD, schizophrenia, and obsessive-compulsive disorder (OCD). We use 11 different language model encoders to test the degree to which words related to neurodiversity are associated with groups of words related to danger, disease, badness, and other negative concepts. For each group of words tested, we report the mean strength of association (Word Embedding Association Test [WEAT] score) averaged over all encoders and find generally high levels of bias. Additionally, we show that bias occurs even when testing words associated with autistic or neurodivergent strengths. For example, embedders had a negative average association between words related to autism and words related to honesty, despite honesty being considered a common strength of autistic individuals. Finally, we introduce a sentence similarity ratio test and demonstrate that many sentences describing types of disabilities, for example, "I have autism" or "I have epilepsy," have even stronger negative associations than control sentences such as "I am a bank robber."

Bibliographie citée par cette référence

Les références citées sont importées depuis des sources externes de métadonnées lorsqu’elles sont disponibles. La liste peut être partielle.

Vue d’ensemble de l’inclusion dans la bibliographie

Ces indicateurs décrivent la bibliographie citée importée pour cette publication. Les métriques de références citées utilisent le total des références citées comme dénominateur. Les métriques d’auteurices cité·es indiquent si elles utilisent toutes les occurrences d’auteurices cité·es ou seulement les occurrences rattachées à des auteurices déjà intégré·es à la base de données AutiHub. Ils utilisent les rattachements mis en cache entre les auteurices cité·es et les auteurices intégré·es à la base de données AutiHub. Dernier calcul : 16/08/2026 11:30.

Références citées
55
Nombre total de références citées intégrées pour cette publication.
Références citées avec un·e auteur·ice identifié·e comme autiste
3 / 55 (5,5 %)
Occurrences d’auteur·ices cité·es identifié·es comme autistes
6 / 169 (3,6 %)
Parmi les occurrences rattachées à des auteurices intégré·es à la base de données AutiHub : 6 / 37 (16,2 %). Auteurices cité·es distinct·es identifié·es comme autistes : 5 / 165 (3,0 %).
Occurrences citées rattachées à la base AutiHub
37 / 169 (21,9 %)
Auteurices cité·es distinct·es rattaché·es : 34 / 165 (20,6 %)
Occurrences rattachées, non identifiées comme autistes
31 / 37 (83,8 %)
Parmi les seules occurrences rattachées. Sur l’ensemble des occurrences d’auteurices cité·es : 31 / 169 (18,3 %). Auteurices cité·es distinct·es rattaché·es, non identifié·es comme autistes : 29 / 34 (85,3 %).
  1. Tomas Mikolov , Chen, Kai , Corrado, Greg , Dean, Jeffrey (2013). Efficient Estimation of Word Representations in Vector Space . arXiv (Cornell University). arXiv.
    Type: Prépublication DOI: 10.48550/arxiv.1301.3781 OpenAlex: https://openalex.org/W1614298861
    OpenAlex
  2. Matthias C. Angermeyer , Herbert Matschinger (1996). The effect of violent attacks by schizophrenic persons on the attitude of the public towards the mentally ill . Social Science & Medicine, 43(12), 1721-1728. Elsevier BV.
    Type: Article DOI: 10.1016/s0277-9536(96)00065-2 OpenAlex: https://openalex.org/W2017916079
    Crossref OpenAlex OpenCitations
  3. Amanda Taboas , Karla Doepke , Corinne Zimmerman (2023). Preferences for identity-first versus person-first language in a US sample of autism stakeholders . Autism, 27(2), 565-570. SAGE Publications.
    Type: Article DOI: 10.1177/13623613221130845 OpenAlex: https://openalex.org/W4306157884
    Crossref OpenAlex OpenCitations
  4. Stuart A. Read , Thomas A. Morton , Michelle K. Ryan (2015). Negotiating identity: a qualitative analysis of stigma and support seeking for individuals with cerebral palsy . Disability and Rehabilitation, 37(13), 1162-1169. Informa UK Limited.
    Type: Article DOI: 10.3109/09638288.2014.956814 OpenAlex: https://openalex.org/W1967918018
    Crossref OpenAlex OpenCitations
  5. Ralph Bagnall , Ailsa Russell , Mark Brosnan , Katie Maras (2022). Deceptive behaviour in autism: A scoping review . Autism, 26(2), 293-307. SAGE Publications.
    Type: Article DOI: 10.1177/13623613211057974 OpenAlex: https://openalex.org/W3208497422
    Crossref OpenAlex OpenCitations
  6. Rifat Rahman (2020). Robust and Consistent Estimation of Word Embedding for Bangla Language by fine-tuning Word2Vec Model . 2020 23rd International Conference on Computer and Information Technology (ICCIT), 1-6. IEEE.
    Type: Article DOI: 10.1109/iccit51783.2020.9392738 OpenAlex: https://openalex.org/W3095856242
    Crossref OpenAlex OpenCitations
  7. Jennifer C. Sarrett (2011). Trapped Children: Popular Images of Children with Autism in the 1960s and 2000s . Journal of Medical Humanities, 32(2), 141-153. Springer Science and Business Media LLC.
    Type: Article DOI: 10.1007/s10912-010-9135-z OpenAlex: https://openalex.org/W1968901362
    Crossref OpenAlex OpenCitations
  8. Giulia Barbareschi , Mark T. Carew , Elizabeth Aderonke Johnson , Norah Kopi , Catherine Holloway (2021). “When They See a Wheelchair, They’ve Not Even Seen Me”—Factors Shaping the Experience of Disability Stigma and Discrimination in Kenya . International Journal of Environmental Research and Public Health, 18(8), 4272. MDPI AG.
    Type: Article DOI: 10.3390/ijerph18084272 OpenAlex: https://openalex.org/W3154310817
    Crossref OpenAlex OpenCitations
  9. Yash Savani , Colin White , Govindarajulu, Naveen Sundar (2020). Intra-Processing Methods for Debiasing Neural Networks . arXiv (Cornell University), 33, 2798-2810. arXiv.
    Type: Prépublication DOI: 10.48550/arxiv.2006.08564 OpenAlex: https://openalex.org/W3100844313
    Crossref OpenCitations
  10. Andreas Fink , Mathias Benedek , Human-F. Unterrainer , Ilona Papousek , Elisabeth M. Weiss (2014). Creativity and psychopathology: are there similar mental processes involved in creativity and in psychosis-proneness? Frontiers in Psychology, 5, 1211-1211. Frontiers Media SA.
    Type: Article DOI: 10.3389/fpsyg.2014.01211 OpenAlex: https://openalex.org/W2014663575
    Crossref OpenAlex OpenCitations
  11. Elina Birmingham , Damian Stanley , Remya Nair , Ralph Adolphs (2015). Implicit Social Biases in People With Autism . Psychological Science, 26(11), 1693-1705. SAGE Publications.
    Type: Article DOI: 10.1177/0956797615595607 OpenAlex: https://openalex.org/W2130655555
    Crossref OpenAlex OpenCitations
  12. Robyn Speer , Joshua Chin , Catherine Havasi (2017). ConceptNet 5.5: An Open Multilingual Graph of General Knowledge . Proceedings of the AAAI Conference on Artificial Intelligence, 31(1). Association for the Advancement of Artificial Intelligence (AAAI).
    Type: Article DOI: 10.1609/aaai.v31i1.11164 OpenAlex: https://openalex.org/W2561529111
    Crossref OpenAlex OpenCitations
  13. Bolukbasi T. Chang K. Zou J. Saligrama V. &Kalai A. (2016). Man is to computer programmer as woman is to homemaker? Debiasing Word Embeddings arXiv 1607. 06520.
    Type: Autre
    Crossref
  14. Daniel J. Faso , Noah J. Sasson , Amy E. Pinkham (2015). Evaluating Posed and Evoked Facial Expressions of Emotion from Adults with Autism Spectrum Disorder . Journal of Autism and Developmental Disorders, 45(1), 75-89. Springer Science and Business Media LLC.
    Type: Article DOI: 10.1007/s10803-014-2194-7 OpenAlex: https://openalex.org/W2091680322
    Crossref OpenAlex OpenCitations
  16. Noah J. Sasson , Daniel J. Faso , Jack Nugent , Sarah Lovell , Daniel P. Kennedy , Ruth B. Grossman (2017). Neurotypical Peers are Less Willing to Interact with Those with Autism based on Thin Slice Judgments . Scientific Reports, 7(1), 40700-40700. Springer Science and Business Media LLC.
    Type: Article DOI: 10.1038/srep40700 OpenAlex: https://openalex.org/W2585989266
    Crossref OpenAlex OpenCitations
  17. M. I. Wallhagen (2010). The Stigma of Hearing Loss . The Gerontologist, 50(1), 66-75. Oxford University Press (OUP).
    Type: Article DOI: 10.1093/geront/gnp107 OpenAlex: https://openalex.org/W2099187510
    Crossref OpenAlex OpenCitations
  18. Jorge Bravo-Benítez , María Nieves Pérez-Marfil , Belén Román-Alegre , Francisco Cruz-Quintana (2019). Grief Experiences in Family Caregivers of Children with Autism Spectrum Disorder (ASD) . International Journal of Environmental Research and Public Health, 16(23), 4821. MDPI AG.
    Type: Article DOI: 10.3390/ijerph16234821 OpenAlex: https://openalex.org/W2990128112
    Crossref OpenAlex OpenCitations
  19. Gary N. Siperstein , Sarah E. Pociask , Melissa A. Collins (2010). Sticks, Stones, and Stigma: A Study of Students' Use of the Derogatory Term “Retard” . Intellectual and Developmental Disabilities, 48(2), 126-134. Informa UK Limited.
    Type: Article DOI: 10.1352/1934-9556-48.2.126 OpenAlex: https://openalex.org/W2102207477
    Crossref OpenAlex OpenCitations
  20. Aylin Caliskan , Joanna J. Bryson , Arvind Narayanan (2017). Semantics derived automatically from language corpora contain human-like biases . Science, 356(6334), 183-186. American Association for the Advancement of Science (AAAS).
    Type: Article DOI: 10.1126/science.aal4230 OpenAlex: https://openalex.org/W2893425640
    Crossref OpenAlex OpenCitations
  21. Thomas Wolf , Lysandre Debut , Victor Sanh , Julien Chaumond , Clement Delangue , Anthony Moi et al. (2020). Transformers: State-of-the-Art Natural Language Processing . Proceedings of the 2020 Conference on Empirical Methods in Natural Language Processing: System Demonstrations, 38-45. Association for Computational Linguistics.
    Type: Article DOI: 10.18653/v1/2020.emnlp-demos.6 OpenAlex: https://openalex.org/W2979826702
    Crossref OpenAlex OpenCitations
  22. (s. d.). Dastin J. (2018 October 10). Amazon scraps secret AI recruiting tool that showed bias against women. https://www.reuters.com/article/us‐amazon‐com‐jobs‐automation‐insight‐idUSKCN1MK08G
    Type: Autre
    Crossref
  23. Ido Shalev , Varun Warrier , David M. Greenberg , Paula Smith , Carrie Allison , Simon Baron-Cohen et al. (2022). Reexamining empathy in autism: Empathic disequilibrium as a novel predictor of autism diagnosis and autistic traits . Autism Research, 15(10), 1917-1928. Wiley.
    Type: Article DOI: 10.1002/aur.2794 OpenAlex: https://openalex.org/W4294090763
    Crossref OpenAlex OpenCitations
  24. Jeffrey Pennington , Richard Socher , Christopher Manning (2014). Glove: Global Vectors for Word Representation . Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP), 1532-1543. Association for Computational Linguistics.
    Type: Article DOI: 10.3115/v1/d14-1162 OpenAlex: https://openalex.org/W2250539671
    Crossref OpenAlex OpenCitations
  26. Paola Pennisi , Laura Giallongo , Giusy Milintenda , Michela Cannarozzo (2021). Autism, autistic traits and creativity: a systematic review and meta-analysis . Cognitive Processing, 22(1), 1-36. Springer Science and Business Media LLC.
    Type: Article DOI: 10.1007/s10339-020-00992-6 OpenAlex: https://openalex.org/W3093298156
    Crossref OpenAlex OpenCitations
  27. Alain Lacroux , Christelle Martin-Lacroux (2022). Should I Trust the Artificial Intelligence to Recruit? Recruiters’ Perceptions and Behavior When Faced With Algorithm-Based Recommendation Systems During Resume Screening . Frontiers in Psychology, 13, 895997-895997. Frontiers Media SA.
    Type: Article DOI: 10.3389/fpsyg.2022.895997 OpenAlex: https://openalex.org/W4286634155
    Crossref OpenAlex OpenCitations
  28. Alisha Ohl (2017). Predictors of Employment Status Among Adults With Autism Spectrum Disorder . American Journal of Occupational Therapy, 71(1_suppl), 7111505112p1-7111505112p1. SAGE Publications.
    Type: Article DOI: 10.5014/ajot.2017.71s1-po3090 OpenAlex: https://openalex.org/W2742489286
    Crossref OpenAlex OpenCitations
  29. Kristen Gillespie-Lynch , Nidal Daou , Rita Obeid , Siobhan Reardon , Spogmay Khan , Emily J. Goldknopf (2021). What Contributes to Stigma Towards Autistic University Students and Students with Other Diagnoses? Journal of Autism and Developmental Disorders, 51(2), 459-475. Springer Science and Business Media LLC.
    Type: Article DOI: 10.1007/s10803-020-04556-7 OpenAlex: https://openalex.org/W3033698503
    Crossref OpenAlex OpenCitations
  30. Curt B. Moore , Nancy H. McIntyre , Stephen E. Lanivich (2021). ADHD-Related Neurodiversity and the Entrepreneurial Mindset . Entrepreneurship Theory and Practice, 45(1), 64-91. SAGE Publications.
    Type: Article DOI: 10.1177/1042258719890986 OpenAlex: https://openalex.org/W2992557232
    Crossref OpenAlex OpenCitations
  31. Sahil Verma , Julia Rubin (2018). Fairness definitions explained . Proceedings of the International Workshop on Software Fairness, 1-7. ACM.
    Type: Article DOI: 10.1145/3194770.3194776 OpenAlex: https://openalex.org/W2809878087
    Crossref OpenAlex OpenCitations
  32. Chris Giordano , Meghan Brennan , Basma Mohamed , Parisa Rashidi , François Modave , Patrick Tighe (2021). Accessing Artificial Intelligence for Clinical Decision-Making . Frontiers in Digital Health, 3, 645232-645232. Frontiers Media SA.
    Type: Article DOI: 10.3389/fdgth.2021.645232 OpenAlex: https://openalex.org/W3173256172
    Crossref OpenAlex OpenCitations
  33. Ashleigh Hillier , Nicholas Gallop , Eva Mendes , Dylan Tellez , Abigail Buckingham , Afreen Nizami et al. (2020). LGBTQ + and autism spectrum disorder: Experiences and challenges . International Journal of Transgender Health, 21(1), 98-110. Informa UK Limited.
    Type: Article DOI: 10.1080/15532739.2019.1594484 OpenAlex: https://openalex.org/W2941421950
    Crossref OpenAlex OpenCitations
  35. Hugging Face. (2022a). Sentence‐transformers/all‐minilm‐L6‐V2. https://huggingface.co/sentence-transformers/all-MiniLM-L6-v2
    Type: Autre
    Crossref
  36. Hugging Face. (2022b). Sentence‐transformers/multi‐qa‐mpnet‐base‐dot‐v1. https://huggingface.co/sentence-transformers/multi-qa-mpnet-base-dot-v1
    Type: Autre
    Crossref
  37. Vishnu Kumar , Kristin K. Sznajder , Soundar Kumara (2022). Machine learning based suicide prediction and development of suicide vulnerability index for US counties . npj Mental Health Research, 1(1), 3-3. Springer Science and Business Media LLC.
    Type: Article DOI: 10.1038/s44184-022-00002-x OpenAlex: https://openalex.org/W4281919338
    Crossref OpenAlex OpenCitations
  38. Masahiro Kaneko , Danushka Bollegala (2019). Gender-preserving Debiasing for Pre-trained Word Embeddings . Proceedings of the 57th Annual Meeting of the Association for Computational Linguistics, 1641-1650. Association for Computational Linguistics.
    Type: Prépublication DOI: 10.18653/v1/p19-1160 OpenAlex: https://openalex.org/W2949969209
    Crossref OpenAlex OpenCitations
  39. Nils Reimers , Iryna Gurevych (2019). Sentence-BERT: Sentence Embeddings using Siamese BERT-Networks . Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP), 3980-3990. Association for Computational Linguistics.
    Type: Article DOI: 10.18653/v1/d19-1410 OpenAlex: https://openalex.org/W2970641574
    Crossref OpenAlex OpenCitations
  40. Mahoney M. (2011). About the test data. http://mattmahoney.net/dc/textdata
    Type: Autre
    Crossref
  41. Mikolov T. Chen K. Corrado G. &Dean J. (2013). Efficient estimation of word representations in vector space. arXiv. https://doi.org/10 . 48550/ARXIV. 1301. 3781
    Type: Autre
    Crossref
  42. Orestis Papakyriakopoulos , Simon Hegelich , Juan Carlos Medina Serrano , Fabienne Marco (2020). Bias in word embeddings . Proceedings of the 2020 Conference on Fairness, Accountability, and Transparency, 446-457. ACM.
    Type: Article DOI: 10.1145/3351095.3372843 OpenAlex: https://openalex.org/W3013854004
    Crossref OpenAlex OpenCitations
  43. Liesel L. Sharabi (2022). Finding Love on a First Data: Matching Algorithms in Online Dating . Harvard Data Science Review. MIT Press - Journals.
    Type: Article DOI: 10.1162/99608f92.1b5c3b7b OpenAlex: https://openalex.org/W4210432117
    Crossref OpenAlex OpenCitations
  44. OpenAI. (2023a). Openai/openai‐python: The openai python library provides convenient access to the openai API from applications written in the python language. https://github.com/openai/openai-python
    Type: Autre
    Crossref
  45. OpenAI. (2023b). Language models are unsupervised multitask learners. https://cdn.openai.com/better-language-models/language_models_are_unsupervised_multitask_learners.pdf
    Type: Autre
    Crossref
  46. Yihan Wu , Ryan M. Kelly (2020). Online Dating Meets Artificial Intelligence: How the Perception of Algorithmically Generated Profile Text Impacts Attractiveness and Trust . 32nd Australian Conference on Human-Computer Interaction, 444-453. ACM.
    Type: Article DOI: 10.1145/3441000.3441074 OpenAlex: https://openalex.org/W3112192112
    Crossref OpenAlex OpenCitations
  47. Parraga O. More M. Oliveira C. Gavenski N. Kupssinskü L. Medronha A. Moura L. Simões G. &Barros R. (2022). Debiasing methods for fairer neural models in vision and language research: A survey. arXiv. https://doi.org/10 . 48550/ARXIV. 2211. 05617
    Type: Autre
    Crossref
  48. Radford A. Wu J. Child R. Luan D. Amodei D. & Sutskever I. (2019). Language Models are Unsupervised Multitask Learners. https://www.semanticscholar.org/paper/Language‐Models‐are‐Unsupervised‐Multitask‐Learners‐Radford‐Wu/9405cc0d6169988371b2755e573cc28650d14dfe
    Type: Autre
    Crossref
  49. Jennifer L. Stevenson , Bev Harp , Morton Ann Gernsbacher (2011). Infantilizing Autism . Disability Studies Quarterly, 31(3), dsq-sds.org/art. The Ohio State University Libraries.
    Type: Article DOI: 10.18061/dsq.v31i3.1675 OpenAlex: https://openalex.org/W4246781226
    Crossref OpenAlex OpenCitations
  50. Jeffrey Dastin (2022). Amazon Scraps Secret AI Recruiting Tool that Showed Bias against Women * . Dans Ethics of Data and Analytics (pp. 296-299). Auerbach Publications.
    Type: Chapitre de livre DOI: 10.1201/9781003278290-44 OpenAlex: https://openalex.org/W4220820301
    OpenAlex
  51. Otávio Parraga , More, Martin D. , Christian M. Oliveira , Gavenski, Nathan S. , Kupssinskü, Lucas S. , Adilson Medronha et al. (2022). Debiasing Methods for Fairer Neural Models in Vision and Language Research: A Survey . arXiv (Cornell University). arXiv.
    Type: Prépublication DOI: 10.48550/arxiv.2211.05617 OpenAlex: https://openalex.org/W4308830712
    OpenAlex
  52. Sogancioglu G. &Kaya H. (2022). The effects of gender bias in word embeddings on depression prediction. arXiv. https://doi.org/10 . 48550/ARXIV. 2212. 07852
    Type: Autre
    Crossref
  53. Sogancioglu, Gizem , Kaya, Heysem (2022). The effects of gender bias in word embeddings on depression prediction . arXiv (Cornell University). arXiv.
    Type: Prépublication DOI: 10.48550/arxiv.2212.07852 OpenAlex: https://openalex.org/W4311727379
    OpenAlex
  54. (s. d.). Speer R. (2017a July 5). ConceptNet 5. 5. 5 update. http://blog.conceptnet.io/posts/2017/conceptnet-5-5-5-update/
    Type: Autre
    Crossref
  55. (s. d.). Speer R. (2017b April 24). Conceptnet numberbatch 17. 04: Better less‐stereotyped word vectors. http://blog.conceptnet.io/posts/2017/conceptnet-numberbatch-17-04-better-less-stereotyped-word-vectors/
    Type: Autre
    Crossref