In 2021, the chief justice, Andrew Cheung Kui-nung, outlined his vision for harnessing technology in the interests of justice. When releasing the judiciary’s annual report, he said that: “In line with our mission to remain a professional and efficient judiciary and move with the times, we have been making a quicker and wider use of technology for enhancing the efficiency of our court operations in various areas.” Those areas included such things as remote hearings in civil and criminal cases, along with “e-bundles at court hearings, e-appointment system, digital evidence and exhibits handling system”.

As an innovator, Cheung will undoubtedly be considering how artificial intelligence (AI) is facilitating the administration of justice in other jurisdictions. In the European Court of Human Rights, for example, it was recently reported that AI, based on a methodology developed jointly by University College London (UCL) and Sheffield and Pennsylvania universities, was able, by automatically analyzing case text using a machine learning algorithm, to predict the verdicts of human rights trials with an accuracy of about 79 percent. The UCL’s head of computer science, Nikolaos Aletras, said AI “could also be a valuable tool for highlighting which cases are most likely to be violations of the European Convention on Human Rights”, which is mind-boggling.

It is, however, not necessary for Cheung to look so far afield, as the Chinese legal system has itself made huge strides in AI usage in recent years, not least through the availability of big data. On his recent trip to Beijing (May 21-25), Cheung not only met the new president of the Supreme People’s Court (SPC), Zhang Jun, but also visited the SPC’s smart court laboratory, the Beijing Financial Court and the Beijing International Commercial Court. Apart from discussing professional judicial work with his hosts, he also considered the use of technology in court operations, which must have been a real eye-opener.

While in Beijing, Cheung also paid a courtesy call on the Supreme People’s Procuratorate, and he was hopefully briefed on how technology is revolutionizing prosecutorial operations. In the Shanghai Pudong People’s Procuratorate, for example, the country’s largest district prosecution office, a machine has been developed that can charge suspects in straightforward cases using AI. Based on a verbal description of the issues, the AI prosecutor can file a charge with an accuracy rate of over 97 percent. This has not only eased the workload of human prosecutors, but also enabled them to concentrate on the more complex cases.

Whereas German prosecutors have been using AI technology for some years for image recognition and digital forensics, given that it facilitates case processing and accuracy, their counterparts in the Chinese mainland have used it since 2016. Chinese prosecutors often use the “System 206”, which can assess the strength of the evidence, the conditions for an arrest, and the suspect’s dangerousness.

The Shanghai AI prosecutor bases its charges on 1,000 “traits”, derived from the case description text it has been given, and which System 206 then assesses. This machine technology was trained by using over 17,000 cases from 2015 to 2020, and it can now determine the charges in Shanghai’s eight most common offenses, including credit card fraud, dangerous driving, and illegal gambling operations.

Although AI can help judges enormously in promoting not only efficiency but also consistent outcomes (including in the sentencing of offenders), it should ultimately be a tool and not a master, and judicial independence must not be usurped in the rush to embrace the latest technology. ... While the judges must always retain ultimate control, there can be no doubt that, if AI technology is responsibly deployed, it can significantly strengthen judicial processes

There are, inevitably, limitations, and an AI prosecutor cannot correctly assess such things as the public interest in bringing a prosecution, or the likely community reaction to the charging of a suspect, or the reliability of a particular witness, which means that the role of the human prosecutor remains paramount.

It is, however, within the judicial system itself that AI’s impact has been most profound, and it is greatly enhancing public justice. According to the SPC’s information center, the smart court SoS (system of systems) now connects the desks of working judges throughout the country. Powered by machine learning technology, the system automatically screens court cases for references, recommends laws and regulations, drafts legal documents and even identifies possible human errors in verdicts.

When the smart court system was first introduced in 2016, it served simply as a database. More recently, however, it has been used more frequently in the decision-making process, and the judges now consult AI in virtually all cases. Indeed, AI has, according to the SPC’s report that was published last year by the Strategic Study of Chinese Academy of Engineering (CAE), cut the average workload of judges by over a third, and it saved citizens 1.7 billion working hours from 2019 to 2021. The report also noted a saving of over 300 billion yuan ($41.8 billion, or HK$327.2 billion) in the same period, which was equal to about half of the total lawyers’ fees in China in 2021.

Whereas, before 2016, each local court maintained its own information system, and judges rarely shared their judgments with their counterparts elsewhere or with Beijing, this, with AI’s advent, has changed. The national smart court system has resulted in every local court converting its documents into a uniform digital format and connecting their databases with the judiciary’s centralized facility in Beijing. The SPC’s report also revealed how AI analyzed nearly 100,000 cases every day across China, and was able to monitor the progress of cases for possible malpractice or corruption. It even helped the higher courts in deciding whether or not to accept an appeal for hearing.

If a court rejects the AI’s recommendation on handling a case, the judge is expected, for the record, to provide reasons. Although some judges may not be happy with this, they may, notwithstanding their reservations over its reliability, accept the AI’s recommendation because of reluctance to challenge the system. It is, therefore, very important to preserve judicial discretion, meaning that judges must always be able to take decisions based on their own evaluation.

Although AI can help judges enormously in promoting not only efficiency but also consistent outcomes (including in the sentencing of offenders), it should ultimately be a tool and not a master, and judicial independence must not be usurped in the rush to embrace the latest technology.

What is also fascinating is the extent to which AI’s influence extends beyond the courtroom. Although verdict enforcement has always been a problem for mainland courts, their new AI capability now helps them to locate and seize the property of a convicted person, and then to put it up for sale by online auction. The smart court can also work with China’s social credit system to ban a person who fails to pay a debt from, for example, using a high-speed train or flying on an aircraft, or even booking into a hotel.

It is clear, therefore, that there is much that Cheung can learn from the technological progress being made in the courts elsewhere in China, although not everything will be of relevance to his own courts. While the judges must always retain ultimate control, there can be no doubt that, if AI technology is responsibly deployed, it can significantly strengthen judicial processes. This, as Cheung undoubtedly appreciates, will also benefit the rule of law.

The author is a senior counsel and professor of law, and was previously the director of public prosecutions of the Hong Kong Special Administrative Region.

The views do not necessarily reflect those of China Daily.