Global mapping of artificial intelligence in Google and Google Scholar

Abstract

The worldwide presence of AI needs to be quantified. This study proposes a descriptive approach and the use of multiple methods and data. An extensive electronic corpus of books was utilized to see the worldwide drift of intellectuals’ minds toward AI and identified related terms, their future trends, and convergence. Using the best bigram proposed by Ngrams Viewer, this study explores the human mind through Google query data, linking popular regions, countries, and related topics to the concept of AI. URL datasets were collected using two popular search engines (SEs), Google and Google Scholar (GS). A URL analysis identified key entities (organizations, institutes, and countries) and their yearly trends. Top-level domains revealed the global web ecology and the annual information growth of AI in SE environments. Information gathered through one approach was fed into the other, revealing a complementary relationship. AI is popular across the globe, and has left traces in many different countries. In this field, GS dominates Google, in relation to the number of sites and domains it includes. Top results reveal the popularity of AI among professionals, artists, programmers, and researchers. The pros and cons of the approaches are also discussed. In addition, this study aims to predict the impact of AI on society, as interpreted through the lenses of well-established theories. The dominance of AI may trap society into aspiring toward an easy life, dependent on intelligent machines. Consistent policies are needed to smooth out future economic cycles in the AI field.

Appendix

Figure 10 shows the results for top-level domains, where the bar labels show the domain percentage and the row labels show the years and TLDs. This figure illustrates the overlapping between SEs. Several top TLDs are exclusive to one SE. AI-related organizations are not attributed to a particular TLD, but top TLDs are few in number. In terms of Domain percentage, the .com TLD is always popular for both SEs, but more so for Google. In terms of ccTLDs, the .in TLD appears only in 2011 with Domain % = 2.8, and then gradually rises each successive year up to 4.3, but only for GS. For India, .in is the ccTLD; owing to the nation’s open policies, it allows unlimited SLDs from India and overseas. In March 2010, the .in TLD had 610,000 domain names, with 60% of registrations from India and the rest from overseas. In March 2016, the number for .in increased to more than 2 million domain names. The Australian ccTLD .au appears consistently from 2004 to 2010, and again in 2016, for GS only. It has recently become more difficult to register as a direct subdomain of .au. The naming rules for .au require registrations under second-level categories that describe a type of entity. For example, .com.au is designed for commercial entities. This rule follows the policy of the U.K. and New Zealand. Among all two-letter countries, the national TLD .uk appears most frequently across all these years. In 2004, 2006, 2008, 2013, 2015, and 2016, Google results contain more AI-related queries, whereas GS has higher results in the remaining years. In March 2012, the .uk TLD was reported to be the fourth most popular ccTLD, after .com, .de, and .net. Figure 10 also presents the combined TLD data for 2004–2016.

Fig. 10

Top TLDs with at least 2.2% domains for the query allintitle: “artificial intelligence”: a results for 2004–2010; b results for 2011–2016 and the combined results for 2004–2016

Figures 11 and 12 show the yearly trends for the top TLDs reported in Fig. 10. AI innovation has been important in Japan since 1980, with the first AI boom continuing until 1987. Here, Japan (.jp) is not as dominant as other AI-supporting countries, such as the U.S. and the U.K. Fig. 13 shows the trend for top countries reported using GTs (Fig. 4) although they are not top for TLDs, except for .ru (Fig. 10, Table 5). Many GT top countries do not have a top TLD % (Fig. 10, Table 5) and vice versa. As previously noted, 19 countries are top-trending countries according to GTs (Fig. 4). The yearly trend for GTs’ countries can be inspected alongside others using ccTLDs (Figs. 11, 12, 13). Six ccTLDs from top countries reported in GTs (Fig. 4) are also at the top in Fig. 10, with their yearly trends shown in Figs. 11 and 12, alongside those of other top countries in Fig. 10. The world population ranking and population percentage for these six countries, as of June 20, 2016, are as follows: Australia (.au, #52, 0.33%), Canada (.ca, #37, 0.49%), India (.in, #2, 17.54%), Japan (.jp, #10, 1.73%), Romania (.ro, #59, 0.27%), and the U.K. (.uk, #22, %0.87). The world population ranking and population percentage of the remaining top 13 GTs countries is not reported. The leading TLDs in Fig. 10 are as follows: Myanmar (.mm, #26, 0.7%), Ethiopia (.et, #13, 1.26%), Kenya (.ke, #29, 0.64%), Malaysia (.my, #43, 0.43%), Nepal (.np, #45, 0.39%), Pakistan (.pk, #6, 2.65%), Peru (.pe, #42, 0.43%), the Philippines (.ph, #12, 1.41%), Russia (.ru, .su, .rf, #9, 2%), South-Korea (.kr, #27, 0.69%), the U.S. (.us, #3, 4.42%), and Vietnam (.vn, #14, 1.25%). Three of these ccTLDs are not found in the URL analysis (Myanmar, Ethiopia, and Peru). The yearly progress of the other ten countries is shown in Fig. 13. AI penetrates small and large countries, in relation to both population and development. For most of the domains, GS trends dominate Google, except in the case of .ca, .net, .com, and .info. In most of the domains the popularity of AI exhibits ups and downs, revealing an upward trend since 2013 in GS, and a decline in GTs in general. A few domains, such as .in, show an upward trend in GS over a longer period. Domains such as .ca, .uk, .edu, .za, .jp, .au, and .edu show a uniform trend since 2004 for both SEs.

Fig. 11

Yearly trends for leading TLDs, also reported in Fig. 9

Fig. 12

Continuing Fig. 11: yearly trends for leading TLDs, also reported in Fig. 9

Fig. 13

Yearly trends for ccTLDs, also reported in GTs

Figures 14–16 are 100% stacked charts, used here to compare domain percentage values to SEs and years (yearly and combined). The study assigned different colors to the ccTLDs, using shades derived from national flags (Fig. 14). These colors are maintained in the subsequent figures, but the approach is not exhaustive, as many countries’ flags share the same colors. TLDs are presented for each year in alphabetical order (left to right), making it easier to appreciate the annual comparison (bottom to top) of these two SEs. In general, gTLDs, namely, .com, .edu, .net, and .org, cover a major area of the web (Fig. 14). As these are old and well-known TLDs, they were expected to be worldwide favorites. The three domains .com, .net, and .org are open TLDs; any entity can use them. The TLD “.edu” is used almost exclusively by colleges and universities in the U.S. It therefore covers a broad area, because U.S. universities dominate the world’s top 100 universities in both ranking and number of universities. The popularity patterns of .edu and .uk are similar; both are becoming increasingly popular in GS every year, but .edu dominates .uk. Meanwhile, the Australian ccTLD .au is more popular in GS, as is the Indian .in. The gTLD .org is more popular in GS, as many organizations still prefer this gTLD. The TLD .info was popular for a few years in Google.

Fig. 14

Top TLDs with a domain % ≥ 2 (annual comparison of the values of domain % to SEs)

Figure 15 shows many new TLDs. The color pattern helps to distinguish those that also appear in Fig. 14, many of which are introduced in GS. Again, the patterning, although not exhaustive, is based on the color of the national flag associated with the ccTLD. Only Columbia .co, Sweden .se, and Russia .ru are introduced in both Google and GS; all of the others appear in GS alone, including Austria .at, China .cn, Taiwan .tw, Cuba .cu, Turkey .tr, Hungary .hu, Greece .gr, South Korea .kr, Croatia .hr, Singapor .sg, Malaysia .my, and Finland .fi. Columbia .co is shown only in 2016 by both SEs. China does not appear in GTs (Fig. 4) as a top country for AI, but these results and the observations of (Yan et al. 2016) show that it is an important country for research in general and AI in particular.

Fig. 15

TLDs with 1 < domain % < 2 (following on from Fig. 14, newly introduced TLDs are patterned)

Many countries are combined in Figs. 15, 16, which follow on from Fig. 14, but with different domain percentage filters. These figures also show that new TLDs (ranked by Domain percentage) are more frequently added to GS than to Google. Argentina .ar is included in Fig. 15 for year 2011 and in Fig. 16 (over many years) for GS. The ccTLD for the European Union .eu also appears in Figs. 15 and 16 for both Google and GS, as does that of Switzerland, .ch. Figure 16 adds countries such as the U.S. .us, South Africa .za, and Egypt .eg for both Google and GS. In Fig. 16, Belgium .bg is shown only in GS, but in both SEs for Domain % ≤ 0.2.

Fig. 16

TLDs with 0.2 < domain % < 1 (following on from Fig. 15, newly introduced TLDs are gradient filled)