China needs urgent AI-specific antitrust rules to curb emerging monopoly risks driven by data control, opaque algorithms and closed tech ecosystems, a former regulator has warned.
China must urgently develop antitrust guidelines for artificial intelligence to counter emerging monopoly threats from data control, algorithm lock-in and closed-source ecosystems, a former Chinese antitrust chief has warned.
"We must accelerate research and formulate antitrust guidelines for the artificial intelligence field,"
Wu Zhenguo, former chief of the Antimonopoly Enforcement Bureau One of the State Administration for Market Regulation, or SAMR, said in a recent article published by LexisNexis, MLex's parent company.
According to data from the Ministry of Industry and Information Technology's press conference last week, the number of AI companies in China exceeded 6,000 and the scale of the core industry is expected to have surpassed 1.2 trillion yuan ($172 billion) in 2025.
Because the industry is developing rapidly, Wu argued that existing rules, based largely on traditional industrial-era contexts, are inadequate for AI's disruptive characteristics.
Wu identified three emerging challenges as AI's innovation model exhibits disruptive characteristics — data dependency, algorithmic black boxes and ecosystem linkages.
First, data ownership remains legally ambiguous while tech giants build "data moats" by controlling key datasets, potentially constituting monopolistic refusal-to-deal conduct, Wu wrote.
Second, core algorithms protected through patents or trade secrets create "technology black boxes" that competitors struggle to overcome. More insidiously, companies may bundle AI services with cloud infrastructure and computing resources or impose exclusivity agreements by leveraging their market position, "forming a chain of blockades across upstream and downstream markets."
Third, giant-led "open source" initiatives often maintain closed-source control over core components, creating a model of "open source for traffic, closed source for profit."
Wu warned that ecosystem monopolies "may not manifest as excessively high patent pricing alone, but rather as control over the interfaces of entire technology ecosystems, discrimination against heterogeneous computing power support, or implicit collusion formed through 'technology alliances'."
To address these challenges, Wu proposed four recommendations.
Most urgently, authorities should develop AI-specific antitrust guidelines that update analytical frameworks when determining relevant markets and dominant position, to incorporate critical factors such as data control, algorithmic influence and user ecosystem dependency.
In addition, clarified standards are needed to identify abusive conduct, to provide clearer guidance on controversial practices such as "refusing to open application programming interfaces, discriminatory algorithm pricing, and training data blockades."
Wu also called for innovative regulatory tools using AI for market-risk scanning and competition simulation; deeper multi-authority governance coordination among antitrust, industrial and judicial authorities; and further participation in international rulemaking on cross-border data flows, algorithm ethics and extraterritorial antitrust jurisdiction.
He urged authorities to prevent any enterprise from leveraging their market position to hinder technological open sourcing or the reasonable flow of data, to maintain the openness of the innovation chain.
Source: MLex
