2023-11-13 19:40:00
Once once more, generative artificial intelligence is causing controversy, now with the National Council of Justice, the CNJ, deciding that it will invest a sentence signed by a federal judge from the 1st Region, made with the help of ChatGPT. The fundamental problem lies in the fact that, in the middle of the argument created by the machine, there was a jurisprudence invented and attributed to the Superior Court of Justice, the STJ.
The sentence, according to the Consultor Jurídica portal, was contested by the lawyer defeated in the case, who noticed the error and contacted the Regional Federal Justice Inspectorate of the 1st Region.
“A case has come to my attention in which a generative AI tool, used as an assistant for drafting a judicial act, presented non-existent precedents as a result of jurisprudential research”, wrote judge Néviton Guedesinspector of the Federal Court of the 1st Region, in a circular.
The judge asks that Justice employees not use tools like ChatGPT for research, especially tools not approved by the Judiciary’s control bodies. Such use, under the magistrate’s responsibility, might affect all civil servants, interns and collaborators involved must compete.
The CNJ, through Resolution 332/2020, granted authorization for the implementation of artificial intelligence within the Judiciary. However, the resolution establishes strict ethical parameters to ensure that the use of these technologies prioritizes the well-being of those under jurisdiction and the guarantee of equitable jurisdiction, stipulating that such tools must, at most, serve as support for judges.
In one specific case, a judge who used ChatGPT argued that it was a “mere mistake” resulting from work overload, mentioning that part of the sentence was prepared by a server. Despite the shelving of the investigation by the Internal Affairs of the 1st Region, the case will now be analyzed by the National Council of Justice.
1699939522
#CNJ #investigate #judges #ruling #ChatGPT #invented #jurisprudence