Por qué el Estado debe regular la Inteligencia Artificial en sanidad: la historia natural de la degradación algorítmica capitalista y las aplicaciones clínicas de la IA
DOI:
https://doi.org/10.11606/issn.2316-9044.rdisan.2025.231445Palabras clave:
Salud digital, Degradación algorítmica, Ética digital, Desigualdades en saludResumen
Objetivo. Describir las características por las que el desarrollo de la inteligencia artificial en aplicaciones clínicas no puede aislarse del proceso general seguido por otros mecanismos de extracción de rentas practicados en plataformas de uso de tecnología digital en otros dominios. Método. Revisión sintética de la literatura académica y periodística reciente que analiza los mecanismos de relación entre concentración de rentas y degradación algorítmica en diversos ámbitos, incluido el sanitario. Resultados. Estos mecanismos conducen inevitablemente a ajustar o afinar los algoritmos y seleccionar los resultados para producir los mejores rendimientos para el capital. La lógica de las corporaciones tecnológicas conduce inexorablemente a un proceso progresivo en el que los mejores intereses de los pacientes deben dejarse de lado siempre que haya buenas razones comerciales. Este proceso es independiente de las buenas intenciones y el altruismo de las start-ups sanitarias. Conclusión. Sin la participación activa del Estado, el ecosistema de generación de soluciones de salud digital se degradará inevitablemente para favorecer la extracción de rentas perjudicando los intereses de los pacientes
Descargas
Referencias
AGGARWAL, Pushkar; PAPAY, Francis. Artificial intelligence image recognition of melanoma and basal cell carcinoma in racially diverse populations. Journal of Dermatological Treatment, v. 33, n. 4, p. 2257–2262, 2021. DOI: DOI: https://doi.org/10.1080/09546634.2021.1944970. Disponível em: https://www.tandfonline.com/doi/full/10.1080/09546634.2021.1944970.
AI SKIN cancer diagnoses risk being less accurate for dark skin. The Guardian, 2021. Disponível em: https://www.theguardian.com/society/2021/nov/09/ai-skin-cancer-diagnoses-risk-being-less-accurate-for-dark-skin-study.
ALI, Aran. The Business of Airbnb, by the Numbers. Visual Capitalist, 06 jan. 2021. Disponível em: https://www.visualcapitalist.com/the-business-of-airbnb-by-the-numbers/.
BAACK, Stefan. The human decisions that shape generative AI: Who is accountable for what?. Mozilla Foundation, 02 ago. 2023. Disponível em: https://foundation.mozilla.org/en/blog/the-human-decisions-that-shape-generative-ai-who-is-accountable-for-what/. Acesso em: 28 abr. 2025.
BELLEMARE, Michel Luc. Techno-capitalist-feudalism. Blacksatin Publishing Limited, 2020.
BENDER, Emily, et al. On the dangers of stochastic parrots: can language models be too big? In: CONFERENCE ON FAIRNESS, ACCOUNTABILITY, AND TRANSPARENCY (FAccT ’21), March 3–10, 2021, Virtual Event, Canada. ACM, New York, NY, USA, 2021 Disponível em: https://dl.acm.org/doi/pdf/10.1145/3442188.3445922.
BOCK, Anna; HSWEN, Yulin. How artificial intelligence will enhance imaging access and analysis. JAMA, v. 331, n. 23, p. 1979-1981, 2024.
BRIN, Sergey; PAGE, Lawrence. The anatomy of a large-scale hypertextual Web search engine. Computer Networks and ISDN Systems, v. 30, p. 107-117, 1998. Disponível em: https://snap.stanford.edu/class/cs224w-readings/Brin98Anatomy.pdf.
BULIK, Beth. Is there a place for pharma in the emerging EHR Market? FiercePharma News, 11 maio 2015.
CAVLAN, O. et al. Real-world evidence: From activity to impact in healthcare decision making. Pharmaceutical and Medical Products. McKinsey & Company, 2017. Disponível em: https://www.paparrizos.org/press/PaparrizosMcKinsey18.pdf.
CHAMPAGNE, David; HUNG, Amy; LECLERC, Oliver. The road to digital success in pharma. Pharmaceutical and Medical Products. McKinsey & Company, ago. 2015.
CFM aponta inconsistências em aplicativo do Ministério da Saúde. Conselho Federal de Medicina, 21 jan. 2021. Disponível em: https://portal.cfm.org.br/noticias/cfm-aponta-inconsistencias-em-aplicativo-do-ministerio-da-saude-2.
CHEN, Irene et al. Ethical machine learning in healthcare. Annu. Rev. Biomed Data Sci., v. 4, p. 123-144, 2021. DOI: https://doi.org/10.1146/annurev-biodatasci-092820-114757. Disponível em: https://www.annualreviews.org/docserver/fulltext/biodatasci/4/1/annurev-biodatasci-092820-114757.pdf?expires=1743713644&id=id&accname=guest&checksum=16B02361C565DAEB497715604658D70E.
DOCTOROW, Cory. The “enshittification” of TikTok. 2023. Disponível em https://www.wired.com/story/tiktok-platforms-cory-doctorow/.
EMILIANA, Cecília. App de Pazuello sugere kit COVID a pacientes com sintomas genéricos. Estado de Minas Nacional, 20 jan. 2021. Disponível em: https://www.em.com.br/app/noticia/nacional/2021/01/20/interna_nacional,1230769/app-de-pazuello-sugere-kit-covid-a-pacientes-com-sintomas-genericos.shtml. Acesso em: 12 jan. 2021.
ENSHITTIFICATION. [s.d.]. Disponível em: https://en.wiktionary.org/wiki/enshittification.
FARZAN, Antonia Noory. A tech company gave doctors free software — rigged to encourage them to prescribe opioids, prosecutors say. The Washington Post, 28 jan. 2020. Disponível em: https://www.washingtonpost.com/nation/2020/01/28/opioid-kickback-software/.
FAUSTINO, Deivisson; LIPPOLD, Walter. Colonialismo digital: por uma crítica hacker-fanoniana. Editora Boitempo, 2023.
HENRY, Terry. Keep pharmaceutical promotion out of doctors’ electronic tools. JAMA, 16 jun. 2021. Disponível em: https://www.ama-assn.org/practice-management/digital/keep-pharmaceutical-promotion-out-doctors-electronic-tools.
KOÇAK, Burak et al. Bias in artificial intelligence for medical imaging: fundamentals, detection, avoidance, mitigation, challenges, ethics, and prospects. Diagn Interv Radiol., v. 31, n. 2, p. 75-88, 2025. DOI: https://doi.org/10.4274/dir.2024.242854. Disponível em: https://www.dirjournal.org/articles/bias-in-artificial-intelligence-for-medical-imaging-fundamentals-detection-avoidance-mitigation-challenges-ethics-and-prospects/doi/dir.2024.242854.
KUWABARA, Masashi et al. Effectiveness of tuning an artificial intelligence algorithm for cerebral aneurysm diagnosis: a study of 10,000 consecutive cases. Sci Rep, v. 13, n. 16202, 2023. Disponível em: https://doi.org/10.1038/s41598-023-43418-x.
MACY Beth. Dopesick: dealers, doctors and the drug company that addicted America. Nova York: Little, Brown & Company, 2018.
MAZZUCATO, Mariana; GERNONE, Fausto. Governments must shape AI’s future. Project Syndicate, 2024 Disponível em: https://www.project-syndicate.org/onpoint/governments-must-shape-ai-future-by-mariana-mazzucato-and-fausto-gernone-2024-04.
MELLO, Michelle; ROSE, Sherri. Denial—artificial intelligence tools and health insurance coverage decisions. JAMA Health Forum, v. 5, n. 3, e240622, 2024. DOI: https://doi.org/10.1001/jamahealthforum.2024.0622.
MILLER, Katherine. Should AI models be explainable? that depends. Stanford HAI, 16 mar. 2021. Disponível em: https://hai.stanford.edu/news/should-ai-models-be-explainable-depends.
MORALES-FORERO, Andres et al. An insight into racial bias in dermoscopy repositories: A HAM10000 data set analysis. J Eur Acad Dermatol Venereol, 14 jun. 2024. DOI: https://doi.org/10.1002/jvc2.477.
O’LEARY, Lizzie. How IBM’s Watson went from the Future of Health Care to sold off for parts. Slate, 31 jan. 2022. Disponível em: https://slate.com/technology/2022/01/ibm-watson-health-failure-artificial-intelligence.html.
REILLY, Tim; STRAUSS, Ilan; MAZZUCATO, Mariana. Algorithmic attention rents: A theory of digital platform market power. Data & Policy, v. 6, 2024. DOI: https://doi.org/10.1017/dap.2024.1.
RÍGOLI, Félix. Inteligência Artificial nos Sistemas de Saúde: Perigos e Promessas para alcançar a saúde para todos In: AITH, Fernando; DALLARI, Analuza Bolivar (Coords.). LGPD na saúde digital. 1. ed. Sao Paulo: Thomson Reuters Brasil, 2022. v. 1, p. 61-80.
ROSS, Casey. Epic’s overhaul of a flawed algorithm shows why AI oversight is a life-or-death issue. Stat News, 24 out. 2022. Disponível em: https://www.statnews.com/2022/10/24/epic-overhaul-of-a-flawed-algorithm/.
SAÚDE lança aplicativo para auxiliar médicos no diagnóstico da Covid-19. ISMEP, 14 jan. 2021. Disponível em: https://www.gov.br/saude/pt-br/assuntos/noticias/saude-lanca-aplicativo-para-auxiliar-medicos-no-diagnostico-da-covid-19.
SAXENA, L. The role of AI-powered decision-making technology in medicare coverage determinations. Special Report,, jan. 2022. Center for Medicare Advocacy. Disponível em: https://medicareadvocacy.org/wp-content/uploads/2022/01/AI-Tools-In-Medicare.pdf
SCHEFFER, Mario; SOUZA, Paulo. A entrada do capital estrangeiro no sistema de saúde no Brasil. Cad. Saúde Pública, v. 38, Sup. 2, e00239421, 2022. DOI: https://doi.org/10.1590/0102-311X00239421.
TAITSMAN, Julie; VANLANDINGHAM, Andrew; GRIMM, Christi. Commercial Influences on Electronic health records and adverse effects on clinical decision-making. JAMA Internal Medicine, v. 180, n. 7, jul. 2020.
TALIA, Domenico. Algorithms that can deny care, and a call for AI explainability. Computer, v. 57, p. 109-112, jul. 2024. Disponível em: https://www.computer.org/csdl/magazine/co/2024/07/10574502/1Y7B8DpFlw4. Acesso em: 28 abr. 2025.
VAROUFAKIS, Yanis. Welcome to the age of Technofeudalism. Wired, 2024. Disponível em: https://www.wired.com/story/yanis-varoufakis-technofeudalism-interview/.
WEN, David et al. Characteristics of publicly available skin cancer image datasets: a systematic review. Lancet Digit Health, v. 4, p. e64-74, 2022.
WONG, Andrew, et al. External validation of a widely implemented proprietary sepsis prediction model in hospitalized patients. JAMA Intern Med., v. 181, n. 8, p. 1065–1070, 2021. DOI: https://doi.org/10.1001/jamainternmed.2021.2626.
YU, Kun-Hsing et al. Medical Artificial Intelligence and human values. N Engl J Med, v. 390, n. 20, p. 1895-1904, 2024. Disponível em: https://www.nejm.org/doi/full/10.1056/NEJMra2214183.
Descargas
Publicado
Número
Sección
Licencia
Derechos de autor 2025 Revista de Direito Sanitário
Esta obra está bajo una licencia internacional Creative Commons Atribución 4.0.
A publicação adota a licença Creative Commons "Atribuição-NãoComercial-CompartilhaIgual 4.0 Internacional" (CC BY-NC-SA). Esta licença permite copiar e redistribuir o material publicado, em qualquer meio ou formato, desde que não se faça uso comercial dos trabalhos publicados e com a devida citação da fonte. Mais informações sobre a licença podem ser encontradas em: <http://creativecommons.org/licenses/by-nc-sa/4.0/>. Ao submeter trabalhos para publicação, os autores aceitam os termos desta licença e concordam em ceder os direitos autorais do manuscrito para a publicação.
