In 2016, AI became part of China’s national technology development program to boost AI research and development and enter formally the race to become a leading AI nation. That China has made tremendous progress highlights a report published by Tsinghua University. According to the report, “China leads the world in AI papers, has become the largest owner of AI patents, has the world’s second largest AI talent pool, and the highest venture investment in AI.” China is running a neck-and-neck race with the United States, followed by countries like Japan and South Korea. Since 2018, however, a debate has also been underway in China about ethical and regulatory questions concerning the use of AI. The National Governance Committee for the New Generation Artificial Intelligence convenes the key Chinese stakeholders from academia, think tanks, businesses, and the government and has recently published its “eight principles for the governance of AI.” The language used and the principles highlighted by the National Governance Committee greatly resembles the language used elsewhere attributing similar concerns to the potentially disruptive nature of intelligent automation.
In fact, a recent study conducted by the Chinese Academy of Sciences reviewed more than 47 AI principle frameworks globally that deal with the question of AI ethics and governance from an international, national, or corporate perspective. The researchers reviewed the AI ethics principles outlined by public and private organizations, such as the OECD, G20, European Commission,USACM, IEEE, Baidu, Tencent, Google, Microsoft, and OpenAI. The study suggests that all organizations emphasize humanity, collaboration, fairness, transparency, privacy, security, safety, and accountability as important topics. Conversely, the assumption that AI has the potential to disrupt modernity and can cause negative consequences towards the economy, society, government, and security is commonly shared. Despite such conceptual overlap, there are also considerable differences in terms of the selection and emphasis of those principles, and none of the reviewed frameworks address all concerns as highlighted by the researchers. Such conceptual gap suggests the need for cooperation as opposed to mere competition. To foster cooperation, a call for “Agile and Comprehensive International Governance of AI” has already been made at another occasion.
Despite the common language, the interpretation of those AI principles can vary profoundly across countries and regions, depending on respective government systems, cultural preferences, and existing local ethics. Thus, despite the conceptual overlap across all frameworks, those differences will primarily determine the actual use of AI in the future. Today, we can already observe the impact of those differences. Despite the risk of polarizing and oversimplifying, in the US, AI is for profit seeking; in Europe, it is used to ensure social cohesion; and in China, it is used to improve harmony and compassion. For example, even though China’s social credit system receives widespread acceptance across the country, it is vehemently criticized in the West because it reveals and clashes with a different value system. Regions, such as Asia, without the history of industrial revolutions are generally less concerned with the risks associated with technological advancement.
The current discourse on AI ethics and governance also omits that ethics cannot simply be defined or changed by fiat and that our existing ethics might contradict newly defined normative demands. Today, we know that our online behavior compromises our privacy, but for the sake of convenience the majority will continue disclosing personal information.
Overall, the debate on AI ethics and governance is a good start, but it needs to include a multidisciplinary and multilateral approach and identify ways to measure the progress of the implementation of those laudable principles. It also need to come to terms with the concept of ethics. For now, it’s vaguely and rather conveniently used to refer to something much more complex, that is social, political, and economic change. If those principles remain as nothing more than ideas, the dominant capitalist spirit and fierce competition among corporations and nations will determine the actual use of AI, leading to major disruptions and unintended consequences.