Chinese researcher Lin Jianxing has warned that artificial intelligence could turn China's social credit system into an automated citizen surveillance and behavioural prediction tool.
Writing in the Journal of Sichuan Normal University, Lin said algorithms risk moving beyond the assessment of past conduct to profiling citizens on anticipated future risks. Under such an arrangement, government bodies could impose restrictions on individuals before any legal violation has taken place.
Lin argued that machine learning models would allow state systems to ingest and analyse far larger volumes of personal data than human administrators can manage. These data points include financial histories, everyday consumer transactions, social ties, and other publicly available digital records. By scanning interconnected digital trails, automated systems could spot patterns that human caseworkers would overlook.
Using these vast datasets, automated algorithms could classify an ordinary resident into a high-risk category, even if that person has never broken a law. Administrative decisions would then be driven by behavioural forecasts rather than concrete actions, leaving targeted individuals facing unexpected travel limits, financial scrutiny, or regulatory checks without being told why.
China's social credit system is frequently misunderstood abroad as a single national database that assigns a mandatory numerical score to every citizen. In reality, the framework is a patchwork of separate mechanisms. These include commercial credit ratings, judicial blacklists for debt defaulters, performance reviews for civil servants, and experimental municipal pilots. In practice, these tools already influence whether individuals and businesses can secure bank loans, purchase high-speed rail tickets, or access certain public contracts.
Stanford University study on municipal ratings
To demonstrate how numerical scoring can malfunction at the local level, Lin pointed to a 2025 Stanford University study that investigated a social credit pilot in a Chinese city. Under that municipal programme, adult residents were allocated a baseline score of 1,000 points.
Individual scores rose or fell based on compliance with 389 specific behavioural rules and criteria for official community rewards. The municipal system tracked everyday civic conduct as well as political engagement, penalising residents for filing petitions and complaints with government offices. Citizens who dropped into the lowest score bracket were subjected to direct police monitoring.
The Stanford researchers discovered that the local pilot produced deep geographic disparities. Rural residents found themselves at a structural disadvantage compared to urban neighbours, because rural communities offered far fewer opportunities to participate in municipal volunteer activities and earn back lost points.
Although the Stanford research found no evidence of an all-encompassing national surveillance network, it confirmed that local pilots could punish citizens for behaviour that did not constitute any criminal offence. Lin warned that integrating modern artificial intelligence into these administrative structures would magnify those risks across entire populations.
Preventing predictive policing of innocent citizens
Lin made clear that his research does not accuse Beijing of actively building a unified, nationwide artificial intelligence rating system. The academic paper presented no evidence that Chinese authorities have begun implementing or designing such a totalised network.
The study is intended instead as a prospective warning for policymakers and technologists before such technologies become entrenched. Lin argued that if state agencies ever adopt algorithmic forecasting for administrative decision-making, clear legal safeguards must be established. Those protections, he wrote, are necessary to prevent the punishment of innocent people and stop excessive intrusion into private life.
The study was conducted with backing from a Chinese state-supported research initiative examining social governance and institutional innovation in public administration. Sichuan Normal University, which published the findings in its academic journal, is a major public research institution located in Chengdu, the provincial capital of Sichuan province in southwestern China.
