New question release: AI “high impact” projects, publications, and skills

Author
Zev Burton
Published
Jun 01, 2022 07:16PM UTC

INFER just launched three new questions within our new topic, Global AI Race: Talent, Research, and Tech (see how we decomposed this topic to develop each forecast question here):


What percentage of Github's “very high impact” AI projects will have EU contributors in 2022?


With over 58 million users and over 300 million repositories, Github is the largest platform for open-source software and collaborations. “Very high impact” projects are defined as projects with more than 100 “forks,” a copy of a repository on which developers can work, and nearly 100,000 repositories fit this description. These projects often include Artificial Intelligence (AI), the subject of our latest strategic question. With a renewed interest in AI from the European Union, INFER is forecasting the percentage of “very high impact” projects to which EU developers have contributed.

Which country or union will have the second most citations of "high impact" AI scientific publications in 2022?


In 2010, there were approximately 162,000 AI publications globally. Eleven years later, that number has more than doubled to over 334,000 publications about AI. Among those publications are “high impact” scientific papers, which OECD defines as having a Field Weighted Citation Impact of over 1.5. The Field Weighted Citation Impact measure, or FWCI, measures the ratio of the actual number of citations received by an output to date and the “expected” number for an article with similar characteristics. (“Expected” refers to the average number of citations for all publications with similar features over the previous three years.) Currently, China authors nearly 40% of all “high impact” scientific publications, but the question remains as to whether the EU, the US, India, or another country will have the second most citations.

How will the U.S. rank in AI skills penetration in 2022?


Even though there are hundreds of thousands of skills that users can add to their LinkedIn profile, few apply to AI. The OECD notes that the top skills comprising the AI skill grouping are machine learning, natural language processing, data structures, computer vision, image processing, deep learning, and several others. These are what is used to measure AI skill penetration, defined as the prevalence of AI skills across occupations within a given country. India has ranked at the top of the list for the past three years, while the United States has shuffled between second and third. The rankings for 2022 have yet to be released, so INFER is forecasting where the United States will rank this year.

These are just the latest questions released on the topic, Global AI Race: Talent, Research, and Tech. Check out more questions on this topic, and be on the lookout for additional questions launching in the coming weeks!

Tip: Mention someone by typing @username