Innovative approach in generative AI

By Liu Jian, Hou Xianli / 09-28-2023 / Chinese Social Sciences Today

On Sept. 21, the 2023 Great Wall Engineering Conference was held in Beijing to explore the innovative development of AI in China. Photo: Sun Zifa/CNSphoto

Recently, generative artificial intelligence has sparked extensive discussions worldwide, raising a number of thought-provoking questions. Will generative AI replace human labor and take away jobs? Will it develop self-awareness? These questions necessitate a thorough analysis of the innovative approach in generative AI. 

Economist Joseph Schumpeter, who popularized the term “creative destruction” in the 20th century, believed that innovation consisted of the upgrading of production technology and the transformation of production methods. American academic Clayton Christensen proposed the theory of “disruptive innovation” in 2003, arguing that technology itself was the innovative outcome of systematic processing of means and ends by humans, and was thus a “finished product” based on the understanding of natural knowledge. Technology inherently possesses objective natural attributes, without the potential for “innovative regeneration” on its own. It can only have “disruptive impact” and bring about “disruptive innovation” when humans expand its use in social contexts. 

It can be seen that “disruptive innovation” emphasizes the integration of technology with humanized nature through human intermediaries, rather than solely focusing on technology itself. Technology is the precondition for successful innovation, which also requires considering business models, value networks, and other factors. 

In the transition from weak AI to strong AI, the disruptive potential of generative AI is often overestimated. In fact, generative AI is not completely replacing innovation and decision-making in business and economic activities as imagined. Since it lacks human practical wisdom, it struggles to autonomously discover specific problems based on macroscopic changes in the real world, which serves as the basis for information retrieval, analysis and integration, as well as innovative decision-making. 

ChatGPT, the widely popular generative AI model, represents an innovative combination of existing engineering technologies in the field. While user experience has been considerably improved, it has not achieved a truly singularity-like breakthrough in the scientific sense. The key natural language processing technologies employed in ChatGPT are relatively mature and have already been proven effective by the academic community in previous years. 

Crossover has now become the trend in the new round of scientific and technological revolution and industrial transformation. Crossover innovation caters to multiple needs within a single product or service, enhancing convenience and comfort in daily life. This phenomenon fully demonstrates human agency, as humans create new wisdom and generate new commercial and market value by recombining seemingly unrelated things according to internal logic and service modes. 

If generative AI is expected to trigger “disruptive innovation,” it is necessary to consider how to better integrate this technology with various application scenarios, and how to cultivate a larger market and explore more needs through crossover innovation, which helps continuously meet new user needs and produce new applications. Generative AI has tremendous potential for application in multiple fields. A technology can only bring about “disruptive innovation” and create greater social and economic benefits when it realizes its full potential across different fields. 

While crossover integration does not necessarily lead to “disruptive innovation,” “disruptive innovations” mostly emerge and develop through crossover integration. This innovation paradigm is called “crossover disruptive innovation” and has become the new norm. Since the “crossover disruptive innovation” constructed by generative AI involves new technologies and new models, its future development should be guided by ideas and values conducive to innovation. 

Therefore, a culture that encourages innovation, tolerates failure, and respects individuality should be fostered, providing a favorable social environment for crossover innovation enabled by generative AI. As innovation continues to deepen and broaden its scope, the crossover integration of multiple innovation systems will be crucial for sustained “disruptive innovation” driven by generative AI. 

An accurate understanding of the generative AI’s innovative approach contributes to appropriately dealing with the relationships between “standardization and development” and “application and ethics,” thereby preventing technological manipulation and leaving room for human liberation and freedom. It also points the way towards “technology for good,” allowing technological development to further benefit human society. 

Liu Jian (lecturer) and Hou Xianli (associate professor) are from the School of Marxism at Heilongjiang University.