Can Judges be Replaced by Machines? The Brazilian Case
The objective of this article is to study the development and use of artificial intelligence (AI) in the Brazilian judiciary in order to assess the likelihood that machines could ultimately replace the judge, and to identify the necessary conditions that might allow for such replacement. To this end, we will examine the relevant concepts related to AI and the scholarship addressing the possibility of replacing the judge with machines in the near future. The methodological procedure used was doctrinal research specific to both new technologies and artificial intelligence. Our conclusion is that AI will reduce the cost of the judicial machinery by allowing many relatively simple and frequently occurring judicial tasks to be resolved more quickly, while leaving ultimate responsibility for judicial decision-making with the judge. Replacement of the judge by automated algorithmic tools would require specific, necessary conditions, even if it were to promote efficiency and cost reduction. These conditions are (i) the adequacy and efficiency of AI generated results, and (ii) the ability of computerized routines to adapt and improve the quality of their application of the law by accommodating and incorporating human corrections of the automated decisions. We also diagnose the philosophical issue that replacing judges with machines might signal a possible “end of interpretation,” even though ongoing human inspection and correction of the computerized systems would be necessary.