While trends in journal articles and conference publications can help us examine trends in late stage R&D activities, examining counts in newspaper articles tied to AI and data science can be equally useful in discerning implementation of these tools and techniques in industries and businesses. The rationale behind these metrics is similar to that for other bibliometric measures. First, given that newspapers make money by selling copies and/or generating funds through selling advertisements, they seek to cover events that are of interest to their readers and will maintain or expand the
number of people reading their articles. Second, implementation of data science and AI related innovations have important implications for employment opportunities, job loss, training required for future jobs, as well as firm profitability, international competitiveness and Government policies.
Based on these two observations, it follows that more articles appear when more firms are adopting these innovations over time. In Figure 1 we present coverage of AI and data science related topics in English language newspapers in Factiva’s database1 and show data confirming the same trends exist when we consider only US and Canadian newspapers (the terms used for the searchers are found in Appendix A). Whether or not we restrict the articles to be both on the topic of AI / data science AND tagged to be related to the region the newspaper is published in, or just focus on any articles referencing AI / data science related terms, it would appear that coverage grew rapidly until 2018, slowed during 2019 and then fell further during the pandemic. This decline in the number of newspaper articles that discuss data science related terms again would suggest that there were fewer innovations and investment opportunities on which to report since the start of the pandemic. Figure 2 shows the coverage of data science topics in Canadian newspapers monthly from January 2020 to May 2021. These more recent trends clearly indicate a rebound consistent with the pattern seen in the LinkedIn Data.
1 https://professional.dowjones.com/factiva/
