Deep Dive: AI Judge Assistant System After One Year — 40% Efficiency Gain, Fairness Debate Continues

In January 2028, China's Supreme People's Court's AI judge assistant system "Tianping" reached one year of operation across 2,800 grassroots courts. System data shows average case processing time has decreased from 87 to 52 days—a 40% efficiency improvement. However, debates over the system's fairness and judicial independence continue to simmer.

Tianping's core functions include intelligent evidence analysis, case precedent recommendations, sentencing suggestions, and automated judgment document generation. For evidence analysis, the system can complete case organization involving 200 pages of evidence materials within 30 minutes, automatically extracting key evidence chains. The case precedent function retrieves rulings from similar cases over the past five years for judge reference.

The efficiency gains have received widespread acknowledgment from grassroots judges. Zhang Hua, presiding judge of the civil division at Beijing Chaoyang District Court, said: "Previously a complex civil judgment document took three days to write. Now AI can generate a draft, and we only need to review and revise—it can be done in half a day."

The focal point of controversy, however, is the sentencing suggestion function. Tianping automatically generates sentencing suggestion ranges based on case facts and historical data. Critics warn this could lead to judges over-relying on algorithmic suggestions, diminishing judicial discretion.

Professor Gu Yongzhong at China University of Political Science and Law's Procedural Law Research Institute noted in a paper: "When AI provides a sentencing suggestion range, whether a judge chooses the lower or upper end is essentially a value judgment. But if judges habitually adopt the midpoint, that value judgment is effectively being made by the algorithm."

Fairness concerns center on training data bias. If historical judgment data contains systematic biases against specific groups, the AI system may learn and amplify those biases. Professor Chen Weidong's research team at Renmin University found that Tianping's compensation suggestions for labor disputes involving migrant workers averaged approximately 15% lower than for similar cases involving white-collar workers.

The Supreme Court responded in December 2027 that a dedicated working group has been formed to conduct a fairness audit of Tianping, with the audit report to be published in Q1 2028. The court emphasized that Tianping is merely an auxiliary tool, with final adjudication authority remaining with judges.

Legal technology companies are actively participating in this space. iFlytek and Baidu have each launched their own judicial AI products, intensifying market competition. China's judicial AI market was estimated at 4.7 billion yuan in 2027, projected to grow to 6.8 billion yuan in 2028.

Internationally, Estonia's AI judge system has operated in small claims for over two years, and the UK is piloting similar AI-assisted sentencing systems. Globally, the integration of AI and justice is accelerating, but ensuring fairness and justice remains an unsolved challenge.