AI research report: China miles ahead

personNews Editor

April 25, 2023

The latest AI annual report, the "Artificial Intelligence Index Report 2023", has just been released, and the data it contains may surprise you. According to the report, China is now miles ahead of all other countries in AI developments. The report, compiled by scientists from Stanford University in the US, spans 386 pages and is an extensive evaluation of data collected worldwide.

Stanford Institute for Human-Centered Artificial Intelligence (HAI)

Nine out of the top ten AI research institutions are from China, with the US's Massachusetts Institute of Technology being the only non-Chinese institution to make the list, coming in at tenth place. The report evaluated the number of AI-relevant publications between 2010 and 2021 to rank the institutions.

Interestingly, universities are becoming less important in AI development, as the report shows that 35 major models for machine learning were published in 2022 - 32 by companies and only 3 by research institutions. This highlights the increasing importance of the private sector in AI research and development.

According to the AI Index Report, nine Chinese universities are among the top 10 AI research institutions.

Although private investment in AI fell for the first time since 2013, the AI Index Report shows that there has been a clear upward trend in private investments in AI in the past ten years. In 2022, private investments in AI were 18 times as high as they were in 2013.

The report also shows that politicians in various countries are dealing with the topic of AI more frequently, with the number of mentions of AI in global legislative processes increasing by a factor of 6.5 between 2016 and 2022.

The researchers found that the approval of AI depends heavily on the country in which you ask. According to an IPSOS survey from 2022, 78% of respondents in China agreed that AI products and services have more advantages than disadvantages, which was the highest approval rating among all the countries examined. In contrast, the researchers found that only 35% of Americans surveyed agreed with the statement.

However, the report also highlights some concerns surrounding AI. The abuse of AI is increasing rapidly, with the number of AI incidents and controversies increasing 26-fold since 2012. Additionally, the report details the impact of AI on the environment, with the training run of the largest open AI language model, "Bloom", causing 25 times as much carbon emissions as a single traveler on a plane trip from New York to San Francisco.

Lastly, the report acknowledges that previous benchmarks for AI will be useless in the future to map further increases in AI capabilities. Therefore, it highlights the importance of new, more comprehensive benchmarking suites such as Google's "BIG-bench" and the "Holistic Evaluation of Language Models" (HELM) developed at Stanford.

The Stanford scientists claim that their AI Index Report is "the world's most credible and comprehensive source for data on and insights into AI". The complete report and the data used are available for free download.

Overall, the "Artificial Intelligence Index Report 2023" offers an extensive evaluation of global AI developments and highlights some of the concerns surrounding AI. With China taking the lead in AI developments, it will be interesting to see how this technology continues to evolve in the coming years.