Chinese tech companies operating within the European Union (EU) are preparing for significant cost increases due to the EU’s sweeping new artificial intelligence (AI) regulations, which will come into force on August 1. Industry experts predict that these rules will drive up assessment and compliance expenses for Chinese enterprises across the EU’s 27 member countries.
Overview of the AI Act
The AI Act, approved by the EU Council in May and passed into law by the European Parliament in March, aims to balance the protection of fundamental rights, democracy and the rule of law with the need for environmental sustainability. It seeks to mitigate the risks associated with high-risk AI while fostering innovation and establishing Europe as a leader in AI technology.
Impact on Chinese tech firms
Hong Kong-based Deyta AI, which specializes in retail analytics software, estimates that complying with the new EU rules will increase its research and development (R&D) and testing costs by 20% to 40%. The higher expenses will include additional documentation, audits, and specific technical measures, said co-founder and CEO Patrick To.
Global race for AI regulation
The enactment and implementation of the EU’s AI rules highlights a global effort to establish AI guardrails, driven by the rise of generative AI (GenAI) services such as OpenAI’s ChatGPT, launched in November 2022. GenAI includes algorithms that can create new content, including audio, code, images, text, simulations, and video, in response to short prompts.
Key provisions of the AI Act
The AI Act categorizes AI technology based on potential risks and impacts. It outlines 12 main headings covering prohibited practices, high-risk systems, transparency obligations, governance, post-market surveillance, information sharing and market surveillance. Additionally, member states are required to establish regulatory sandboxes and real-world testing at the national level.
Compliance challenges
Non-compliance with the AI Act can result in hefty fines – up to 35 million euros (US$38 million) or 7% of the offending firm’s total worldwide annual turnover, whichever is higher. Dayta AI’s Tu emphasized the importance of complying with mandates related to the quality, relevance and representativeness of training data, which will ultimately enhance the performance and fairness of their solutions.
Comparison with Chinese regulations
The user rights-centric approach of the AI Act imposes strict limits on personal data use, unlike Chinese regulations, which emphasize technological advancement and alignment with government priorities. According to Beijing’s GenAI rules, implemented on August 15 last year, firms must adhere to core socialist values and prohibit content that threatens political and social stability.
Regional developments in AI legislation
Other jurisdictions in Asia are also developing AI regulations. South Korea has drafted its own “Act on the Framework for Promoting the AI Industry and Establishing Trustworthy AI”, which is currently under review. According to Alex Roberts, partner at Linklaters and head of the firm’s China TMT group, there is a trend in the Asia-Pacific region for governments to adopt elements from the EU’s AI regulation when drafting their own legislation.
Conclusion
The EU’s comprehensive AI rules, effective on August 1, are set to serve as a global blueprint amid growing efforts by governments to regulate AI technology. While these regulations present compliance challenges for Chinese tech firms, they also provide an opportunity to enhance the quality and fairness of AI solutions through diligent data management and adherence to strict standards.