1. Responsabilidade algorítmica, personalidade eletrônica e democracia.
- Author
-
Amadeu da Silveira, Sérgio
- Subjects
- *
ARTIFICIAL neural networks , *MACHINE learning , *PRECAUTIONARY principle , *ARTIFICIAL intelligence , *CIVIL law - Abstract
This text analyzes the argument about algorithmic systems of machine learning and deep learning, especially about artificial neural networks, presented as inscrutable, unfathomable and incomprehensible. It exposes the criticism and the possible ineffectiveness of transparency to support the explainability of Artificial Intelligence solutions. It exhibits the incompatibility of algorithmic opacity with its democratic use in socially important applications. Explains the claim of inscrutability and the impossibility of interpretation the operations performed by the deep learning algorithms allows the construction of a new legal figure, the electronic person, as occurred with the Recommendations to the Commission on Civil Law Rules on Robotics of European Parliament, in 2017. Unable to measure the consequences in the case the proposition be adopted, the text presents the precautionary principle to support the convenience and necessity of regulations and risk restrictions of a technology that extends the power of platforms and large corporations. [ABSTRACT FROM AUTHOR]
- Published
- 2020