Whether chatbot-generated opinion can be trusted: Application of the hearsay rule of the Criminal Justice Act 2003 and inspiration for China's evidence law reform

Abstract

As part of its criminal justice reform, the UK Ministry of Justice issued a policy paper on the use of software-generated evidence in criminal proceedings. In response, this article consistently uses the term “chatbot-generated opinion,” while employing “chatbot testimony” metaphorically to compare chatbot responses in court to human testimony given in answer to a judge's questions. The article explores whether chatbot-generated opinion evidence can be trusted, focusing on the hearsay rule under the UK Criminal Justice Act 2003 and its implications for reforming China's evidence law. Contrary to prevailing views that exclude such evidence due to the lack of cross-examination, the article argues that chatbot opinions should not be directly accepted as testimony. It further explains that virtue jurisprudence offers an appropriate framework for identifying indicators of justified belief. Since a chatbot is incapable of having a moral motivation, this prevents judges from justifiably believing a chatbot's statement qua testimony. Introducing such an evidence analysis approach in China is significant. China's Confucian ethics and virtue jurisprudence highlight moral responsibility and motivation, providing a valuable foundation for the ongoing reform of China's evidence law and overcoming the limitations of its predominant objectivist approach to proof.