|
How AIGC shapes the “AI divide”: Generation mechanisms and bridging pathways
2024, 42 (10):
2017-2027.
Generative artificial intelligence (AIGC) is being integrated into social life and is participating in knowledge production. It also creates new digital inequalities, and scholars are concerned about deepening the digital divide. This article explores the theoretical context of artificial intelligence and its impact on society and tries to answer how AIGC shapes the "AI divide" in the context of the development of Digital China. At the technical level, the "AI divide" involves three dimensions: data, algorithms, and computing power. Data sets serve as the foundation of AIGC, similar to soil. The scale parameters, source channels, content types, and even languages of the data sets invisibly shape the "AI divide". The algorithm model consists of self-supervised learning and human feedback reinforcement learning. It is similar to the roots and stems of a plant, which determines the vitality of AIGC to bloom and bear fruit. Computing power is AIGC's information infrastructure, connecting the real world (chips) and the virtual world (data and algorithms), determining the processing speed of computing operations and the depth of integration in the digital industry. As modern technology becomes increasingly dominant in shaping our understanding of the world and history, social stratification theory can help us understand the impact of the "AI divide". The first level involves the divide among individuals, specifically between ordinary users, in accessing and utilizing AI technology. This includes language barriers, paywalls, practical skills, willingness to use, perception of benefits, and the desire to learn. The second aspect pertains to human-machine differentiation, which, on the surface, refers to the distinction between humans and autonomous AI agents. However, at its core, it represents the divide between the majority who lack technological initiative and the minority who wield it due to their control over digital resources. It refers to using digital resources by countries, regions, or individuals to participate in the development, share knowledge, and overcome differences in capabilities for autonomous growth. To bridge the "AI divide", it is important to establish an independent and self-sufficient technological innovation system. We should also explore and optimize application scenarios to make the most of this system. China needs to take the lead in AI development in the future and promote the construction, transformation, and application of AIGC resources by different enterprises, scientific research institutions, and industry organizations. It must also establish a fair and standardized platform governance ecosystem and create a good technology atmosphere. Equality should be embedded into AIGC's technology process, including algorithm design, training data selection, model generation, and optimization. This will give people of different countries, ethnic groups, and classes the right to participate in scientific development. Finally, China should plan more diverse ecological alliance forms and participate in cooperation platforms under multilateral frameworks such as the United Nations, the World Trade Organization, and Asia-Pacific Economic Cooperation. China should transform games and competitions into consultation and cooperation, connect wider upstream and downstream resources, and try to build a global AI governance community. The article proposes the concept of the "AI divide", which deepens the third-generation digital divide theory in artificial intelligence and has significant practical implications for understanding the development of digital China and the economic and social transformation.
|
|