Between continuities and ruptures: the representation of scientists and science based on images generated by ChatGPT
DOI:
https://doi.org/10.11606/issn.2316-9125.v29i1p127-146Keywords:
Artificial intelligence, Representation, Stereotypes, Scientist, ScienceAbstract
Stereotypical images of scientists and science may interfere with the trust of the public and movements toward or away from the scientific field. Given the increasing use of artificial intelligence in contemporary times, we aim to analyze images generated by ChatGPT regarding “scientist” and “science” to find continuities and ruptures in the maintenance of racial, gender, and age stereotypes, as well as those that point to the exact and biological sciences as a reference model. Overall, results suggest the reaffirmation of the scientist as a white man, as well as of exact and biological sciences as central in society. They also indicate changes in scientific practice, shifting from the individual to the collective, bringing an optimistic view regarding the inclusion of young scientists and the ease of access to hypertechnological resources.
Downloads
References
AMARASEKARA, Inoka; GRANT, Will J. Exploring the YouTube science communication gender gap: A sentiment analysis. Public Understanding of Science, Thousand Oaks, v. 28, n. 1, p. 68-84, 2019. DOI: https://doi.org/10.1177/096366251878665
ASH, Elliott et al. Visual representation and stereotypes in news media. SSRN, New York, p. 1-26, 2021.
BENDER, Emily et al. On the dangers of stochastic parrots: can language models be too big? Proceedings of the 2021 ACM Conference on Fairness, Accountability, and Transparency, [s. l.], p. 610-623, 2021. DOI: https://doi.org/10.1145/3442188.3445922
BETKER, James et al. Improving image generation with better captions. OpenAI, [s. l.], p. 1-19, 2023. DOI: https://cdn.openai.com/papers/dall-e-3.pdf
BIANCHI, Federico et al. Easily Accessible text-to-image generation amplifies demographic stereotypes at large scale. Proceedings of the 2023 ACM Conference on Fairness, Accountability, and Transparency, [s. l.], p. 1493-1504, 2023. DOI: https://doi.org/10.1145/3593013.3594095
CENTRO DE GESTÃO E ESTUDOS ESTRATÉGICOS (CGEE). Percepção pública da C&T no Brasil – 2019. Resumo executivo. CGEE, [s. l.], 2019. Disponível em: https://www.cgee.org.br/web/percepcao
CHAKRAVERTY, Snehashish; MAHATO, Nisha Rani; SAHOO, Deepti Moyi.. McCulloch-Pitts neural network model. In: Concepts of Soft Computing, New York, p. 167-173, 2019. DOI: https://doi.org/10.1007/978-981-13-7430-2_11
CHAMBERS, David Wade. Stereotypic images of the scientist: the draw a scientist test. Science Education, [s. l.], v. 67, n. 2, p. 255-265, 1983. Disponível em: https://acesse.dev/He0JM. Acesso em: 1 jul. 2024.
CHIMBA, Mwenya Diana; KITZINGER, Jenny. Bimbo or boffin? Women in science: an analysis of media representations and how female scientists negotiate cultural contradictions. Public Understanding of Science, Thousand Oaks, p. 609-624, 2010. Disponível em: https://www.researchgate.net/publication/51108862_Bimbo_or_boffin_Women_in_science_An_analysis_of_media_representations_and_how_female_scientists_negotiate_cultural_contradictions. Acesso em: 1 jul. 2024.
COZMAN, Fabio Gagliardi. Inteligência Artificial: uma utopia, uma distopia. Teccogs: Revista Digital de Tecnologias Cognitivas, São Paulo, v. 17, n. 17, p. 32-43, 2018. DOI: https://doi.org/10.23925/1984-3585.2018i17p32-43
DAMASCENO, Daniel et al. Injustiça epistêmica e reafirmação de estereótipos: a representação do cientista no Fantástico e Domingo Espetacular durante a pandemia da Covid-19. Contracampo, Niterói, n. 1, v. 43, p. 1-17, 2024. DOI: https://doi.org/10.22409/contracampo.v43i1.61118
FEUERRIEGEL, Stefan et al. Generative AI. Business & Information Systems Engineering, New York, v. 66, n. 1, p. 111-126, 202. DOI: https://doi.org/10.1007/s12599-023-00834-7
FLICKER, Eva. Between brains and breasts ‒ women scientists in fiction film: on the marginalization and sexualization of scientific competence. Public Understanding of Science, Thousand Oaks, v. 12, n. 3, p. 307-318, 2003. DOI: https://doi.org/10.1177/0963662503123009
FRANKENBERG, Ruth (ed.). Displacing whiteness: Essays in social and cultural criticism. Durham: Duke University Press, 1997.
FRASER, Kathleen C.; KIRITCHENKO, Sevtlana; NEJADGHOLI, Isar. Diversity is not a one-way street: pilot study on ethical interventions for racial bias in text-to-image systems. Proceedings of the 14th International Conference on Computational Creativity, [s. l.], 2023. Disponível em: https://computationalcreativity.net/iccc23/papers/ICCC-2023_paper_97.pdf. Acesso em: 1 jul. 2024.
GAMKRELIDZE, Tamari; ZOUINAR, Moustafa; BARCELLINI, Flore. Working with Machine Learning/Artificial Intelligence systems: workers’ viewpoints and experiences. Proceedings of the 32nd European Conference on Cognitive Ergonomics, [s. l.], p. 1-7, 2021. DOI: https://doi.org/10.1145/3452853.3452876
HALL, Stuart. Cultura e representação. Rio de Janeiro: Apicuri/PUC-Rio, 2016.
Idem. Representation: cultural representations and signifying practices. Londres: Sage, 1997.
HARAWAY, Donna. Saberes localizados: a questão da ciência para o feminismo e o privilégio da perspectiva parcial. Cadernos Pagu, São Paulo, n. 5, p. 7-41, 1995. Disponível em: https://periodicos.sbu.unicamp.br/ojs/index.php/cadpagu/article/view/1773
HARDING, Sandra. Strong objectivity: A response to the new objectivity question. Synthese, New York, v. 104, n. 3, p. 331-349, 1995. Disponível em: https://link.springer.com/article/10.1007/BF01064504
HAYNES, Roslynn D. The scientist in literature: images and stereotypes-their importance. Interdisciplinary Science Reviews, [s. l.], v. 14, n. 4, p. 384-398, 1989. DOI: https://doi.org/10.1179/isr.1989.14.4.384
KALOTA, Faisal. A primer on generative artificial intelligence. Education Sciences, v. 14, n. 2, 2024. DOI: https://doi.org/10.3390/educsci14020172
KING, Morgan. Harmful biases in artificial intelligence. The Lancet Psychiatry, London, v. 9, n. 11, p. e48, 2022. DOI: https://doi.org/10.1016/S2215-0366(22)00312-1
LAMBRECHT, Anja; TUCKER, Catherine. Algorithmic Bias? An Empirical Study of Apparent Gender-Based Discrimination in the Display of STEM Career Ads. Management Science, [s. l.], v. 65, n. 7, p. 2966-2981, 2019. DOI: https://doi.org/10.1287/mnsc.2018.3093
LARSON, Jeff et al. How We Analyzed the Compas Recidivism Algorithm. ProPublica, [s. l.], 2016. Disponível em: www.propublica.org/article/how-we-analyzed-the-compas-recidivism-algorithm. Acesso em: 1 jul. 2024.
LOBO, Paula; CABECINHAS, Rosa. The negotiation of meanings in the evening news: towards an understanding of gender disadvantages in the access to the public debate. International Communication Gazette, Thousand Oaks, v. 72, n. 4-5, p.339-358, 2010. DOI: https://doi.org/10.1177/1748048510362611
LOPES, Maria Margaret. Sobre convenções em torno de argumentos de autoridade. Cadernos Pagu, São Paulo, n. 27, p. 35-61, 2006. DOI: https://doi.org/10.1590/S0104-83332006000200004
LUCY, Li; BAMMAN, David. Gender and representation bias in GPT-3 generated stories. Proceedings of the 3rd Workshop on Narrative Understanding, Long Beach, p. 48-55, 2021. DOI: https://doi.org/10.18653/v1/2021.nuse-1.5
MARTINEZ, A. R. Representation matters: theorizing health communication from the flesh. Health Communication, Abingdon, v. 38, n. 1, p. 184-190, 2021. DOI: https://doi.org/10.1080/10410236.2021.1950293
MASSARANI, Luisa; CASTELFRANCHI, Yurij; PEDREIRA, Anna Elisa. Cientistas na TV: como homens e mulheres da ciência são representados no Jornal Nacional e no Fantástico. Cadernos Pagu, São Paulo, n. 56, p. 1-34, 2019. DOI: https://doi.org/10.1590/18094449201900560015
MASSARANI, Luisa et al. O que os jovens brasileiros pensam da ciência e da tecnologia. Resumo executivo. Rio de Janeiro: INCT-CPCT, 2019.
MCCARTHY, John et al. A proposal for the dartmouth summer research project on artificial intelligence. August 31, 1955. AI Magazine, Washington, D.C., v. 27, n. 4, p. 12-14, 2006. DOI: https://doi.org/10.1609/aimag.v27i4.1904
NEIVA, Silmara Cássia Pereira Couto et al. Perspectivas da ciência brasileira: um estudo sobre a distribuição de bolsas de pesquisa em produtividade do CNPq ao longo do ano de 2019. Revista Interdisciplinar Científica Aplicada, Blumenau, v. 16, n. 3, p. 51-71, 2022. Disponível em: https://portaldeperiodicos.animaeducacao.com.br/index.php/rica/article/view/18090. Acesso em: 1 jul. 2024.
RAMESH, Aditya et al. Zero-shot text-to-image generation. Proceedings of the 38th International Conference on Machine Learning, Long Beach, v. 139, p. 8821-8831, 2021. Disponível em: https://proceedings.mlr.press/v139/ramesh21a.html?ref=journey-matters. Acesso em: 1 jul. 2024.
REZNIK, Gabriela; MASSARANI, Luisa Medeiros; RAMALHO, Marina; MALCHER, Maria A.; AMORIM, Luis; CASTELFRANCHI, Yurij. Como adolescentes apreendem a ciência e a profissão de cientista? Revista Estudos Feministas, Florianópolis, v. 25, n. 2, p. 829-855, 2017. DOI: https://doi.org/10.1590/1806-9584.2017v25n2p829
REZNIK, Gabriela; MASSARANI, Luisa Medeiros; MOREIRA I de C. Como a imagem de cientista aparece em curtas de animação? História, Ciências, Saúde, Manguinhos, v. 26, n. 3, p. 753–777, jul. 2019. DOI: https://doi.org/10.1590/S0104-59702019000300003
SALINAS, Abel et al. The unequal opportunities of large language models: examining demographic biases in job recommendations by ChatGPT and LLaMA. Proceedings of the 3rd ACM Conference on Equity and Access in Algorithms, Mechanisms, and Optimization, Long Beach, n. 34, 2023. DOI: https://doi.org/10.1145/3617694.3623257
SARKAR, Sujan. Uncovering the ai industry: 50 most visited ai tools and their 24B+ traffic behavior. Writerbuddy, [s. l.], 2023. Disponível em: https://writerbuddy.ai/blog/ai-industry-analysis. Acesso em: 1 jul. 2024.
SCHIENBINGER, Londa. O feminismo mudou a ciência? Bauru: Edusc, 2001.
SCHULMAN, John et al. Introducing ChatGPT. OpenAI, [s. l.], 2022. Disponível em: https://openai.com/blog/chatgpt#OpenAI. Acesso em: 1 jul. 2024.
SCHWARCZ, L. K. M. O espetáculo das raças: cientistas, instituições e questão racial no Brasil: 1870-1930. São Paulo: Companhia das Letras, 1993.
SICHMAN, Jaime Simão. Inteligência Artificial e sociedade: avanços e riscos. Estudos Avançados, São Paulo, v. 35, n. 101, p. 37-50, 2021. DOI: https://doi.org/10.1590/s0103-4014.2021.35101.004
TAMBKE, Erika. Mulheres Brasil 40º: os estereótipos das mulheres brasileiras em Londres. Espaço e Cultura, Rio de Janeiro, n. 34, p. 123-150, 2013. Disponível em: https://www.e-publicacoes.uerj.br/espacoecultura/article/view/12744. Acesso em: 1 jul. 2024.
TEIXEIRA, Pedro. ChatGPT reforça estereótipos sobre mulheres brasileiras: magras, bronzeadas e com acessórios coloridos. Folha de S. Paulo, São Paulo, 6 mar. 2024. Disponível em: https://www1.folha.uol.com.br/tec/2024/03/chatgpt-reforca-estereotipos-sobre-mulheres-brasileiras-magras-bronzeadas-e-com-acessorios-coloridos.shtml. Acesso em: 1 jul. 2024.
VASWANI, Ashish et al. Attention Is All You Need. Proceedings of the 31st International Conference on Neural Information Processing Systems, Long Beach, p. 6000-6010, 2017.
VYAS, Darshali A. et al. Hidden in plain sight: reconsidering the use of race correction in clinical algorithms. New England Journal of Medicine, New England, v. 383, n. 9, p. 874-882, 2020. DOI: https://doi.org/10.1056/nejmms2004740
ZACK, Travis et al. Assessing the potential of GPT-4 to perpetuate racial and gender biases in health care: a model evaluation study. The Lancet Digital Health, London, v. 6, n. 1, p. e12-e22, 2024. DOI: https://doi.org/10.1016/s2589-7500(23)00225-x
Downloads
Published
Issue
Section
License
Copyright (c) 2024 Luiz Felipe Fernandes Neves, Amanda Medeiros, Luisa Massarani
This work is licensed under a Creative Commons Attribution 4.0 International License.
I authorize the publication of the submitted article and soon the copyrights to the magazine, in the printed and electronic version, if it is approved after the evaluation of the reviewers.
I understand that readers may use this article without prior request, provided the source and authorship are mentioned. Readers are not authorized to use this article for reproduction, in whole or in part, for commercial purposes.