What can Brussels learn from China's generative AI regulation?
The article highlights a critical divergence in global approaches to regulating generative AI, emphasizing that current legal frameworks are struggling to keep pace with rapid technological deployment. While companies rush to embed AI into products despite known flaws, regulators are attempting to manage risks through disparate enforcement actions and new legislative drafts. This creates a complex landscape where the lack of unified standards forces users and developers into uncertain legal territories, underscoring the urgent need for coherent governance strategies that balance innovation with fundamental rights. A stark contrast emerges between the EU and China regarding data and copyright protections. China’s draft law imposes stringent requirements, mandating explicit consent for personal data and prohibiting the use of infringing copyright material for training models. In contrast, the EU’s AI Act adopts a softer approach, relying on transparency summaries and self-assessments rather than proactive prohibitions. This discrepancy is particularly relevant to open data discussions, as it questions whether the EU’s permissive stance inadvertently facilitates the unauthorized scraping of publicly available datasets, potentially undermining ethical data practices and the rights of content creators. Furthermore, the text critiques the EU’s reliance on self-regulation and its silence on truth and misinformation, warning of potential enforcement failures similar to those seen with the GDPR. The article argues that without strict consent-based models for both personal data and copyrighted content, open data ecosystems risk being corrupted by illegal or unethical data sourcing. For open data advocates, this situation emphasizes the necessity of advocating for regulations that protect individual autonomy and intellectual property, ensuring that open data principles do not come at the expense of legal compliance and human rights.
Source: euronews.comPublished on 2023-05-24