Inteligência artificial e neurotecnologias em diagnósticos e tratamentos médicos: o colapso do consentimento na saúde algorítmica

Abstract

This dissertation examines the use of artificial intelligence (AI) systems and neurotechnologies in medical diagnosis and treatment, focusing on how their technical characteristics – especially algorithmic opacity, the “double black box” between neural networks and brain circuits, and structural data biases – strain the protection of the fundamental rights to health, equality, data privacy, and mental integrity. The central aim is to show that the convergence of opacity and algorithmic discrimination progressively erodes patient autonomy and leads to the functional collapse of the traditional informed consent model as designed for analog medicine. The dissertation adopts a legal-dogmatic and interdisciplinary approach, grounded in bibliographic research, analysis of regulatory documents, and comparative study across Brazil, the European Union, and the United States. Drawing on paradigmatic cases of diagnostic AI and brain-computer interfaces, it first maps the clinical and operational potential of these technologies (increased accuracy, therapeutic personalization, and service efficiency) and, second, the risks arising from opacity, bias, and the possibility of manipulating clinical decision-making and patient behavior through digital nudges and dark patterns. The proposition defended is that, in an intelligent and neurotechnological healthcare environment, patient consent alone is no longer a sufficient instrument of protection, requiring a transition to a paradigm of effective self-determination anchored in “structural guardianship”. This protection entails positive duties of transparency, explainability, bias auditing, neurodata governance, and the recognition of specific safeguards for the mental sphere (neuro-rights), as well as a regime of shared responsibility among developers, health professionals, institutions, and regulators. It concludes that only a robust normative and institutional architecture – one that combines consent, fundamental rights, and preventive mechanisms of technological governance – can reconcile innovation in AI and neurotechnologies with patient dignity and equality in the age of algorithmic health care.