Artigo Original
MEDICAL EXPERTISE IN THE AGE OF ARTIFICIAL INTELLIGENCE: BIAS, POST-TRUTH, AND THE RISK OF COHERENT ERROR IN THE CONSTRUCTION OF EXPERT EVIDENCE
Como citar: Oliveira HBD. MEDICAL EXPERTISE IN THE AGE OF ARTIFICIAL INTELLIGENCE: BIAS, POST-TRUTH, AND THE RISK OF COHERENT ERROR IN THE CONSTRUCTION OF EXPERT EVIDENCE. Persp Med Legal Pericia Med. Vol. 11, 2026; 260533.
https://dx.doi.org/10.47005/260533
Recebido em 09/04/2026
Aceito em 21/05/2026
The authors report no conflict of interest.
MEDICAL EXPERTISE IN THE AGE OF ARTIFICIAL INTELLIGENCE: BIAS, POST-TRUTH, AND THE RISK OF COHERENT ERROR IN THE CONSTRUCTION OF EXPERT EVIDENCE
Resumo
INTRODUÇÃO: A mediação algorítmica do raciocínio técnico desloca o problema do viés do indivíduo para um sistema híbrido humano-tecnológico, no qual a IA influencia seleção, hierarquização e interpretação das evidências, alterando a construção da prova pericial  METODOLOGIA: Revisão narrativa estruturada com busca em bases biomédicas e documentos institucionais, incluindo estudos experimentais, revisões sistemáticas, metanálises e artigos conceituais, com distinção entre níveis de evidência  FUNDAMENTAÇÃO TEÓRICA: O viés em IA é distribuído ao longo do ciclo de vida, podendo ser reforçado pela interação humano-máquina, enquanto evidências recentes demonstram a possibilidade de manipulação estrutural do conhecimento interno dos modelos  RESULTADOS: Predominam estudos observacionais com risco de viés elevado, além da demonstração de que modelos podem incorporar informações incorretas de forma persistente e coerente, mantendo desempenho aparente  DISCUSSÃO: Propõe-se o conceito de erro coerente, caracterizado por coerência lógica sustentada em premissas epistemicamente comprometidas, inserido no contexto da pós-verdade e amplificado pela mediação algorítmica  CONCLUSÃO: O desafio central é epistemológico, exigindo vigilância crítica sobre a integridade das premissas, reconhecimento dos limites da coerência como critério de validade e preservação do julgamento independente do perito 
Palavras Chave: Pericia medica; inteligencia artificial; vies; epistemologia; prova pericial; pos-verdade.
Abstract
INTRODUCTION: Algorithmic mediation transforms technical reasoning by influencing the selection, prioritization, and interpretation of evidence, reshaping the construction of expert testimony  METHODS: A structured narrative review was conducted using biomedical databases and institutional documents, including experimental studies, systematic reviews, meta-analyses, and conceptual articles, with attention to levels of evidence  THEORETICAL FRAMEWORK: AI bias is distributed throughout its lifecycle and reinforced through human-machine interaction, while recent evidence demonstrates the possibility of structural manipulation of internal model knowledge  RESULTS: The literature is predominantly observational with high risk of bias, and models can incorporate incorrect biomedical information persistently and coherently while maintaining apparent performance  DISCUSSION: The concept of coherent error is proposed, defined as logically consistent reasoning built upon epistemically compromised premises, amplified in a post-truth context and by algorithmic mediation  CONCLUSION: The central challenge is epistemological, requiring critical vigilance over the integrity of premises, recognition of the limits of coherence as a validity criterion, and preservation of independent expert judgment 
Keywords (MeSH): Medical expertise; artificial intelligence; bias; epistemology; expert evidence; post-truth.
References
1. Hasanzadeh F, Josephson CB, Waters G, Adedinsewo D, Azizi Z, White JA. Bias recognition and mitigation strategies in artificial intelligence healthcare applications. NPJ Digit Med. 2025 Mar 11;8(1):154.
2. Vicente L, Matute H. Humans inherit artificial intelligence biases. Sci Rep. 2023 Oct 3;13(1):15737.
3. Pearson J, Dror IE, Jayes E, Whordley GR, Mason G, Nightingale S. Examining human reliance on artificial intelligence in decision making. Sci Rep. 2026 Feb 5;16(1):5345.
4. Han T, Nebelung S, Khader F, Wang T, Müller-Franzes G, Kuhl C, et al. Medical large language models are susceptible to targeted misinformation attacks. NPJ Digit Med. 2024 Oct 23;7(1):288.
5. Ledley RS, lusted LB. Reasoning foundations of medical diagnosis; symbolic logic, probability, and value theory aid our understanding of how physicians reason. Science. 1959 Jul 3;130(3366):9-21.
6. Shortliffe EH. Computer-based medical consultations: MYCIN. New York: Elsevier; 1976. 263 p.
7. Esteva A, Kuprel B, Novoa RA, Ko J, Swetter SM, Blau HM, et al. Dermatologist-level classification of skin cancer with deep neural networks. Nature. 2017 Feb 2;542(7639):115-18.
8. Rajpurkar P, Irvin J, Zhu K, Yang B, Mehta H, Duan T, et al. CheXNet: radiologist-level pneumonia detection on chest X-rays with deep learning. arXiv [preprint]. 2017 [cited 2026 May 11]. Available from: https://arxiv.org/abs/1711.05225
9. Topol EJ. High-performance medicine: the convergence of human and artificial intelligence. Nat Med. 2019 Jan;25(1):44-56.
10. McKinney SM, Sieniek M, Godbole V, Godwin J, Antropova N, Ashrafian H, et al. International evaluation of an AI system for breast cancer screening. Nature. 2020 Jan;577(7788):89-94.
11. Singhal K, Azizi S, Tu T, Mahdavi SS, Wei J, Chung HW, et al. Large language models encode clinical knowledge. Nature. 2023 Aug;620(7972):172-80.
12. Lee P, Bubeck S, Petro J. Benefits, Limits, and Risks of GPT-4 as an AI Chatbot for Medicine. N Engl J Med. 2023 Mar 30;388(13):1233-9.
13. McIntyre L. Post-truth. Cambridge (MA): MIT Press; 2018. 240 p.
14. Arendt H. Truth and politics. New York: Viking; 1968. Chapter 1, Truth and politics.
15. Social Issues Research Centre, Royal Society, Royal Institution of Great Britain. Guidelines on science and health communication. 1st ed. Oxford: SIRC; 2001. 8 p.
16. Graham P. How to disagree [Internet]. Weymouth: PaulGraham. [2008] – [cited 2026 Mar 10]. Available from: http://www.paulgraham.com/disagree.html
17. Conselho Nacional de Justiça. Mecanismos de inteligência artificial poderão apoiar decisões judiciais em saúde [Internet]. Brasília (DF): CNJ. [2025] – [cited 2026 Mar 20]. Available from: https://www.cnj.jus.br/mecanismos-de-inteligencia-artificial-poderao-apoiar-decisoes-judiciais-em-saude/
18. Barros TTD, Motta MVD, Muñoz D. Inteligência artificial na medicina legal e perícia médica brasileira: inovação tecnológica em meio à ausência de regulamentação específica. Persp Med Legal Pericia Med. 2026;11:260208.
