China’s soft regulatory approach aimed to provide impetus for investments!!!
On the 15th of August in the year 2023, China enacted a novel legal framework designed to regulate generative AI. This legislation, representing the latest in a series of regulatory endeavors targeting diverse facets of AI, stands out as an internationally pioneering initiative, for it singularly addresses the domain of generative AI. This legal enactment introduces fresh constraints on entities providing generative AI services to the public, delving into the realms of both the data utilized in training and the content generated as outputs.
Notwithstanding the imposition of new restrictions on commercial enterprises offering generative AI services, an analysis of the evolution of the draft legislation in conjunction with the broader tech policy milieu might inadvertently suggest that China is poised to relax its staunch approach to AI regulation. Observers have expeditiously noted that the final iteration of the generative AI regulations has been considerably diluted in comparison to the earlier draft that was released for public commentary. Provisions mandating swift rectification of illegal content within a three-month timeframe and the assurance of “truthful and accurate” training data and output have been conspicuously excised. It is also expressly stipulated that these regulations exclusively apply to public-facing generative AI systems. Moreover, a new provision has been incorporated, emphasizing the equitable weighting of development and innovation alongside the imperatives of system security and governance.
Nonetheless, these indicia can be somewhat misleading in comprehending the future trajectory of AI policy in China. It is pertinent to acknowledge that alterations to the text of draft AI regulations following a consultation period are not uncommon. For instance, a draft AI regulation concentrating on recommender systems in 2021 witnessed the removal of explicit provisions pertaining to anti-discrimination safeguards.
While the attenuation of the generative AI regulations may arguably surpass that of prior initiatives, it is essential to recognize the ongoing efforts to institute effective AI regulation, encompassing the early draft stages of a prospective comprehensive AI legislation. These initiatives are emblematic of a consistent drive to fortify the country’s AI regulatory framework.
Similarly, the nomenclature “tech crackdown” has been loosely applied to policies entailing diverse government agencies, targets, and justifications. While some measures, such as investigations into technology corporations, were largely reactive and appear to have tapered off, the formulation of robust AI regulations has long been a policy aspiration of the Chinese government, likely to persist. In aggregate, these factors signify that China continues to refine the delicate equilibrium between innovation and control in its approach to AI regulation, rather than embarking on a substantive relaxation.
The groundbreaking strides taken by China in introducing AI regulations, coupled with the history of reactionary measures curbing tech enterprises, might exert a dampening effect on industry outcomes in the short term. This challenge is further compounded by the repercussions of U.S. export controls on semiconductor technology to the Chinese AI sector, which has compelled companies to devise workarounds amid scarcity of potent microchips. Although China has sought to bolster its AI sector through financial support, access to computing resources, and ministerial restructurings aimed at fostering domestic innovation, the fruits of these endeavors remain uncertain.
Nevertheless, notwithstanding the immediate impact on China’s AI sector, the implementation of regulations designed to oversee AI is imperative for mitigating the perils inherent in these technologies. Such regulations, with the practical mechanisms they mandate, serve to ameliorate harm to individuals and preserve social stability, particularly through requisites such as watermarking AI-generated content, a potent tool against the proliferation of misinformation and disinformation.
By contrast, the relatively laissez-faire approach adopted by the United States leaves it potentially ill-prepared to address these risks. Furthermore, AI governance tools bolster China’s ambitions of assuming global leadership in the AI domain, particularly through the formulation of international standards, which could bestow a competitive advantage.
In sum, it is improbable that China’s fundamental approach to AI regulation will undergo a substantial transformation, even amid concurrent economic uncertainties. While stringent regulatory measures may present economic challenges in the short term, they remain indispensable for averting harm to individuals, upholding social stability, and securing international dominance in regulatory frameworks over the long tearm.