
Japan’s Supreme Court plans a fiscal 2027 examination of generative artificial intelligence (AI) in civil proceedings, following an earlier conclusion that the technology should serve as an assistant to judges.
That conclusion came from a test conducted in January and February by a research group comprising six mid-career judges from the Tokyo and Osaka district courts.
Mock trial records formed the basis of the exercise.
The judges identified possible benefits in using AI to prepare summaries and organise points of contention.
They also found that the technology could reinforce judges’ biases by omitting important facts or disputed issues and could produce answers that appeared convincing despite being incorrect.
Funding for the next stage appears in the court’s fiscal 2027 budget request, announced on Monday (August 31).
The request includes about 60 million yen in AI-related expenses, marking the first time the Supreme Court has included such costs in a budget request.
It also contains about 38,743 million yen for digitising trial procedures, including criminal and domestic cases.
The demonstration will draw its material from lawsuits that have already concluded.
After judges and other officials enter documents into an AI system, it will be tested on tasks such as producing rough drafts of record summaries and lists of points of contention.
Under another part of the plan, the court will develop a system to prevent the disclosure of litigants’ personal information and companies’ trade secrets.
Prolonged civil trials have become a problem in recent years.
Masayuki Nagata, a senior official at the Supreme Court’s digital policy planning office, said: “Using AI is expected to streamline and shorten trials further.”
Nagata added: “We hope to present the results of the demonstration test within the next fiscal year.”
[Copyright The Jiji Press, Ltd.]