Artificial Intelligence papers and their destiny

Have you ever wondered how many scientific articles are published every year? In 1980, about 1 million articles were published. Since then, this figure has been growing at an impressive rate — at 2014 it exceeded 7 million. At the same time, computer science articles are among the leaders of this trend. Over 334k articles on AI were released in 2021.

Is there a way to understand what role a particular paper has played in the development of AI? The metrics that answer this question use citation as the main unit of account. The number of papers that mentioned the article in question is used either directly, as an absolute value, or divided by the sum of all citations of a given scientific field —SNIP (Source-Normalized Impact per Paper).

An obvious feature of these metrics is their isolation within the academic world. Altmetric partially circumvents this limitation by including not only citations but also the number of mentions of the article in the media, as well as the number of downloads in various services. However, this still leaves us in a world of circumstantial evidence.

So still, is there a way to understand which article is destined to get lost in a thousand other unremarkable works, and which one will change the world around us forever? One of the natural ways to answer this question is to collect and analyze examples of the use of various AI technologies on tasks that solve real problems for humanity, especially in areas not directly related to the world of AI.

What are the consequences of the “Artificial Intelligence butterfly effect” (butterflAI effect) for the world around us? What has changed and continues to change in various areas of life around since the term Artificial Intelligence was first used in scientific circles?

Let’s try to figure it out in the upcoming posts.