China has deployed the strictest AI control system in the world, with direct State supervision over each algorithm and language model that is put into operation, as detailed in an extensive study published in Internal Comparative Legal Guides by four local experts.
Why is it important. The Asian giant is building a regulatory framework that contrasts with the European one. The European Union hopes to analyze the risks and self-regulation prioritizing the protection of rights. China, on the other hand, requires prior authorizations and continuous supervision of each AI-based service, in addition to facing enormous public investments.
The panoramic. The Chinese plan is based on three pillars:
- Prior control. Each algorithm and AI model must register with the authorities within a maximum of ten days after its launch.
- Continuous supervision. Companies must label all AI-generated content and maintain strict control throughout the life of each service, not just at launch.
- social protectionism. There are specific regulations for vulnerable groups, such as children and the elderly.
The Chinese system is already giving results. Tech giants such as Microsoft and HP have registered their algorithms, while authorities have shut down AI services that flouted the rules.
In 2024, the Chongqing CAC ordered the immediate closure of “Lingxiang Zhiwen AI” for operating without security assessments, and “Kaishanhou AI Writing Master” was sanctioned for generating illegal content without carrying out mandatory reviews.
Go deeper. The Chinese strategy differs radically from the European AI Law. While the EU categorizes apps according to their level of risk and relies on business self-regulation, China combines strong state control with massive investment to accelerate development.
Any foreign company that wants to operate in China must deploy its servers there and request authorizations through local entities. It is something that we have already seen in the past and that now extends to AI… and that contrasts with the European digital single market.
In detail. The plan also includes strict rules on the labeling of AI-generated content, protection of personal data and intellectual property, along with an ethical framework with specific requirements on algorithmic bias.
The combination of regulation and investment is bearing fruit: Chinese companies such as Baidu and SenseTime are developing AI models that compete with OpenAI and Google, while state control ensures that the technology serves national interests.
In techopiniones | The strange rapprochement between Trump and Apple has an explanation: Samsung
Featured image | techopiniones with Midjourney