Only 2.5% of AI papers and sustainability related researchers: shameful data

“We analyzed papers on artificial intelligence empowering sustainable development and found that out of 10 million English published papers related to ICT (Information and Communication Technology) and artificial intelligence, only 2.5% were related to sustainable development. This is a very shameful data, and for AI researchers like me, I believe there are many aspects that can be done better.”

Recently, at the Zhongguancun Forum Future Artificial Intelligence Pioneer Forum, Zeng Yi, Director of the Long Term Artificial Intelligence Research Center, expressed the above exclamation. Empowering various industries with generative artificial intelligence is one of the hottest topics at present. However, Zeng Yi believes that the direction of artificial intelligence development is not only to make money, but also for enterprises and scholars to have a broader vision. The government should promote applications related to artificial intelligence that cannot be profitable in the short term, and promote important issues for the whole society, especially the goal of global sustainable development.


Among the 2.5% of papers related to sustainable development mentioned above, Zeng Yi mentioned that the vast majority are the combination of artificial intelligence and health, and artificial intelligence and high-quality education. Why are these two fields? “Although these two fields are very important, what is more crucial is that they are very profitable, and the other 15 fields are not very profitable.” Zeng Yi said that the artificial intelligence industry and scholars are not very concerned about other issues, which poses significant problems for the development of the world.

Perhaps it may not necessarily help everyone earn much money, but artificial intelligence can play an empowering role in ecological relationships and cultural inheritance. Zeng Yi stated that one of the research conducted by the team is the use of artificial intelligence to enhance the understanding of symbiotic ecological relationships. “This world may not necessarily be centered around humans. There are many types of animals and plants around the world, and how they relate to humans. We can now answer this question based on artificial intelligence technology, draw such symbiotic maps, and show everyone the relationship between humans and other animals.”

In addition, artificial intelligence can also empower cultural heritage. When we go global, we see more of the fragmentation between cultures because there are some differences in values. “When applying for World Cultural Heritage and Natural Heritage, we need to emphasize their uniqueness, but it is generative artificial intelligence that tells us that in reality, the connection between these natural cultural heritage sites far exceeds their uniqueness.”

Artificial intelligence can indeed empower various industries, but there is also a potential issue of how to address labor substitution. One of the United Nations Sustainable Development Goals is to promote full and productive employment and decent work for all. However, how to avoid the potential impact of AI on labor employment in the future and explore new ways to better unleash human creativity and wisdom are risks that need to be avoided while promoting sustainable development.

In addition, artificial intelligence will also have negative impacts on sustainable development, such as development and zero carbon issues. “Everyone is currently pursuing big data and computing power, but in the future, we must use greener energy and lower carbon efficient ways to achieve artificial intelligence technology, so that we can contribute to the goal of zero carbon while transforming towards intelligence.”

Zeng Yi mentioned that the United States has done the best in AI governance globally, and at the same time, the United States accounts for 63% of the world’s reports on artificial intelligence risks, while China accounts for 9% of the world’s reports on risks. From 2022 to 2023, the number of risks associated with artificial intelligence has increased more than 10 times globally, and predictions show that the risks associated with artificial intelligence in 2024 will be 10 times higher than in 2023.

“The capacity building of artificial intelligence not only includes building infrastructure for data and knowledge platform software and hardware, as well as application services, but also includes the ability to govern AI, cultivate talent, and enhance public AI literacy.” Zeng Yi believes that China’s artificial intelligence will definitely go global, but there are two problems in going global. The first is the problem of global development, and the second is the problem of international governance. In global sustainable development, we still have many goals that are not done well enough.

At this forum, the Forward Artificial Intelligence Research Center also released an Action Plan, proposing to promote sustainable development through specific actions. The Action Plan points out that while promoting sustainable development, attention should be paid to potential negative impacts, such as using green energy and low-carbon and efficient artificial intelligence technologies as much as possible, and paying attention to the potential impact on labor employment while using artificial intelligence empowerment.

“This is a world composed of humans and ecology, and artificial intelligence still makes many mistakes that people do not make, so it still has many risks. Humans and the world can be assisted by AI, but should not be dominated by AI. Artificial intelligence does not need to be ubiquitous, and we should be responsible, steadily developing, and moderately using it. Only in this way can artificial intelligence develop steadily.” Zeng Yi said.