Why Do Men in Their 40s and 50s Push AI So Hard? A Comparison with the Famicom, Pagers, and the Internet

The heaviest users are not the loudest advocates

There is a common impression that the people who talk most passionately about “AI is the future” are men in their 40s and 50s. No statistic was found that measures this impression by age. But if we ask whether these people are the heaviest users, the statistics say no.

In the 2025 survey by Japan’s Ministry of Internal Affairs and Communications (MIC; January–February 2026, 1,236 valid responses in Japan), the share of people who had used generative AI was 78.2% at ages 20–29, 49.0% at 40–49, and 44.2% at 50–59. At ages 15–19 it was 91.7%. In use, young people lead, and people in their 40s and 50s sit below the middle.

The pattern is similar in the US. Bick and colleagues surveyed Americans aged 18–64 in 2024 and found that the share using generative AI at work was 28.7% at 18–29, 31.0% at 30–39, 30.5% at 40–49, and 17.5% at 50–64. People in their 40s are level with the younger groups but do not exceed them.

The question can then be restated. Why is the advocacy so loud when the use is not the highest? This note lines up the information revolutions this generation has lived through since childhood and looks there for the reasons. No study was found that measures the loudness of advocacy itself by age, so the reasons given here are hypotheses built from verified facts.

Scope and method

Generation: people aged 40–50 in 2026, that is, born 1976–1986. Following the request’s word “ojisan” (middle-aged men), the note focuses on men, though most statistics combine men and women.

Revolutions compared: to the three in the request (video games, pagers and mobile phones, the internet), two missing candidates were added: the mobile internet (i-mode) and the smartphone. The reason is given in the section that maps the revolutions onto ages.

Sources: 22 sources were used: Japanese official statistics and white papers (MIC, Science Council of Japan), official corporate material (Nintendo, NTT DoCoMo), research on technology generations and attitude formation, the sociology of expectations, empirical studies of generative AI use, and the social and economic history of Japan’s information society. The full text of every source was retrieved, and figures and claims were checked against the relevant passages. For sources where only the abstract could be read (Humlum & Vestergaard 2025, Nakagomi et al. 2025, Kushida 2011, Jorgenson & Motohashi 2005, Vogel 2013), only what the abstract says was used, and the text says so.

Limits: “advocacy” is a different behavior from use, and no statistic measures it by age or gender. The note therefore stops at building hypotheses by combining statistics on use and position, research on formative years, and the social history of each revolution.

Mapping five revolutions onto ages

How old were people born 1976–1986 when each revolution arrived?

The Famicom went on sale on 15 July 1983, and 19.35 million units were sold in Japan (Nintendo). Pager subscriptions peaked at more than about 10 million in fiscal 1995 and began to decline from June 1996 (NTT DoCoMo Technical Journal, Figure 1)¹. Individual internet users grew from 11.55 million (9.2%) at the end of 1997 to 27.06 million (21.4%) at the end of 1999 and 47.08 million (37.1%) at the end of 2000 (MIC White Paper 2010)². MIC’s 2019 White Paper writes that Windows 95, released in 1995, “is said to” have been a major trigger for the internet’s spread to the general public, and that users of the earlier PC communication services grew to 5.73 million in 1996. i-mode launched on 22 February 1999 with the slogan “from a phone for talking to a phone for using” (NTT DoCoMo). Household smartphone ownership rose to 9.7% at the end of 2010, 29.3% at the end of 2011, and 49.5% at the end of 2012 (MIC White Paper 2013).

The reason for adding ④ and ⑤ is in the second column from the right. i-mode (④) made the phone an internet device in Japan; it spread in a different place from ③ (the pocket) and from a different group (secondary-school students). The smartphone (⑤) was the first information revolution this generation received as adults, so they received it from a different position than ③ and earlier. Without these two, generative AI (⑥) would have no counterpart to compare with as a revolution received after the formative years.

Did the Famicom take children’s outdoor play away?

Revolution ① in the request comes with the view that the Famicom moved children’s play from outdoors to indoors. This view needs correcting on timing.

The 2013 recommendation of the Science Council of Japan’s subcommittee on children’s developmental environments writes that children’s outdoor play time fell below indoor play time around 1965. The same recommendation says that in the twenty years from about 1955 to about 1975 (the “first change”), outdoor play time halved, from 3.2 to 1.8 hours for boys and from 2.3 to 1.0 hours for girls, and states that “it was television that took outdoor play away from children.” In the next twenty years (to about 1995, the “second change”), outdoor play time fell further to 0.62 hours, and the recommendation names the rise in lessons such as cram schools and sports clubs as one cause. The Famicom’s release falls in the middle of this second change.

Tsuruyama and colleagues, surveying elementary school children in Toyama in 2005, likewise reported that “games and PCs” (38.3%) and “outdoor play” (36.8%) were nearly tied as the favorite form of play, and attributed the decline in outdoor play to a combination of fewer places to play, no spare time because of cram schools and lessons, and fewer playmates because of the falling birth rate. So the Famicom did not start the move of play indoors. It fits the sources better to say that the Famicom put new content, “a screen you can control yourself,” into play that had already moved indoors because of television and lessons.

This correction also matters for the comparison with ⑥. What the Famicom left behind was less a change of place than the feel of moving things on a screen yourself. Many boys born 1976–1986 entered information technology from this side, the side of the one who operates.

Whose were pagers and mobile phones?

Pagers began as call devices that adults were made to carry for work. The NTT DoCoMo Technical Journal describes the feeling of users in the era when pagers were used only for emergency calls as being “made to carry” them and “tied down.” Once pagers could send numbers and kana, the market spread to groups that had not used them before.

At the center of those groups were young women. Ito and Okabe (2005), drawing on interviews with high-school and university students near Tokyo (winter 2000) and communication diaries collected from July to February up to 2003, write that messaging had been strongly associated with girls, especially kogyaru (high-school “gals”), since the pager era. In the same passage, they contrast the male otaku associated with video games and computers with media-savvy girls associated with communication devices such as pagers and mobile phones³. The university students (aged 18–21) whose diaries Ito and Okabe collected were, counting back from the survey period, born roughly 1981–1985, the second half of this note’s generation. According to the authors, many heavy users of that generation used pagers in middle and high school in the mid-1990s, moved through PHS, and switched to the mobile internet in the late 1990s.

Two things can be read from ②. One is that the communication revolution was perceived as having spread, as a tool for fun and socializing, from the girls’ side. The other is that each time, adults worried about the young. Those born 1976–1986 were on the side being worried about.

The internet arrived in the middle of the formative years

For this generation, the internet (③) sits differently from the other revolutions. The idea of technology generations explains the difference.

Sackmann and Winkler (2013) revisited the idea of technology generations developed in Germany in the early 1990s. In that definition, a technology generation forms from the birth cohorts whose formative years (roughly ages 15–25) coincide with a new technology reaching 20% of households. Because household internet diffusion in Germany passed 20% in 2000, those born after 1980 form the internet generation, and those born 1964–1979 are the preceding computer generation.

In Japan, individual internet use passed 20% at the end of 1999 (21.4%). Because the indicators differ (households versus individuals), and the authors themselves caution that their grouping rests on European diffusion patterns and would need adjusting for other regions, this is only a guide, but the timing is almost the same as in Germany. At that time, those born 1976–1986 were 13–23, entering or in the middle of their formative years. Applying the German grouping directly, the first half of this generation (born 1976–1979) is the last of the computer generation and the second half (born 1980–1986) is the first of the internet generation. Either way, this generation saw “the internet in the middle of changing society” at the age when views settle.

The idea that the formative years shape later views has support from another line of research. Krosnick and Alwin (1989), using US election panel studies, found results supporting the impressionable years hypothesis: attitudes are highly susceptible to change in late adolescence and early adulthood, and susceptibility drops sharply afterward. The competing hypothesis, that people grow gradually more resistant to change throughout life, was rejected. That study dealt with political attitudes, not attitudes toward technology. Applying it to views of technology is this note’s inference.

For this generation, ③ and ④ were “revolutions that changed everything, seen while young.” The request’s note on ③, “everything changed,” can be read as a way of putting this experience. They received the smartphone (⑤) when they were already working. Generative AI (⑥) arrived 15–20 years after their formative years.

Four hypotheses about why they push AI

From these facts, the note proposes four hypotheses about why people in their 40s and 50s push AI strongly. None has been confirmed by research that measures advocacy itself.

Hypothesis 1: The scene of “everything changing,” seen in the formative years, becomes a template

This generation watched, in their formative years, the internet change society. That experience can become a template for reading a new information technology as “this will change everything in the same way.” Sackmann and Winkler write that technology generations are shaped in their formative years by interactions with technological interfaces and purposes, and that this influences access to and use of later new technologies. Generative AI is easy to fit into that template. Bick and colleagues report that generative AI has spread faster than the PC and the internet; including use outside work, adoption two years after the first mass-market product exceeded that of the PC and the internet. Seeing it as “faster than the thing I saw before” may strengthen the template further.

Hypothesis 2: Using past futures in present claims about the future

Brown and Michael (2003), treating technological expectations sociologically, distinguished two activities. One is recalling how the future of a technology was once described (retrospecting prospects); the other is incorporating those recalled futures into present-day future-making (prospecting retrospects).

If people in their 40s and 50s say “it was the same with the internet” when they push AI, that can be read as recalling the futures described in the era of ③ (and the fact that many came true) and using them as grounds for claims about AI’s future. This generation heard the internet’s future described when they were young and saw it actually happen as adults. So they hold past futures as “prophecies that came true.” Younger generations grew up with ③ as a given and do not share the same memory.

Hypothesis 3: The memory of Japan going first and being left behind

i-mode (④) let Japan use the internet on mobile phones before the rest of the world. But that lead did not spread to the world. Kushida (2011), in the abstract, says Japan’s telecommunications sector became decoupled from global markets, producing a “Galapagos effect” in which winning in an isolated domestic market led to losing in global markets, and attributes this not only to misguided technological choices but to competitive dynamics shaped by politics and regulation. Vogel (2013), in the abstract, writes that Japan lost its edge in areas of its greatest competitive strength, electronics and especially ICT hardware, declining in market share, exports, and profits. Jorgenson and Motohashi (2005), in the abstract, report that the share of Japanese GDP devoted to IT investment rose sharply after 1995, while productivity growth in the non-IT sector lagged far behind the United States.

Those born 1976–1986 used the i-mode lead as secondary-school and university students, and watched, as newly working adults, the 2000s and 2010s in which Japanese electronics lost its strengths. This experience can remain as a lesson: “Even if you have it first, you get left behind if you ride it wrong.” With generative AI too, the share of Japanese companies with a policy to use it (active and limited combined) is 68.9%, lower than 82.8% in the US (MIC 2025 survey). The strength of the advocacy is likely mixed with an urgency of “don’t miss the boat this time.”

Hypothesis 4: The first revolution met on the deciding side

From ① to ④, this generation was the young side that used new technology first and that adults worried about. The Science Council recommendation says excessive video game play should be curbed, and pagers and mobile phones spread together with adult discourse lamenting youth’s moral decline (Ito & Okabe). They received the smartphone (⑤) as working adults.

With generative AI (⑥), this generation is for the first time on the side that decides adoption at work. MIC’s 2025 corporate survey took as respondents managers or above at companies with 10 or more employees, and 86.4% in Japan answered that their company used generative AI in at least one task. The voice of “our company pushes AI” is likely to come from people in positions like those of the respondents. Use (individual usage rates) and advocacy (statements pushing adoption within an organization) are measured in different places. A division in which young people use AI as individuals and people in their 40s and 50s push it as an organization may produce the gap from the opening: advocacy is loud even though use is not the highest.

Why “ojisan”, that is, why men?

The sources explain only half of the male skew.

The half they explain is that generative AI use itself skews male. In Bick and colleagues’ US survey, use at work was 7.5 percentage points higher for men than for women. For PCs in 1984, by contrast, use was 6 points higher for women, which the authors attribute to the many women in secretarial and administrative jobs. At least in the case of the PC, the gender gap depended less on the technology itself than on which occupations it entered first. Humlum and Vestergaard (2025), surveying 18,000 workers in 11 occupations in Denmark, report in the abstract that women were 16 percentage points less likely than men to have used ChatGPT for work. In Japan, Nakagomi and colleagues (2025) report from a survey of 13,367 internet users that men used generative AI about 1.8 times as much as women (according to Chiba University’s press release).

The other half is the difference in childhood entry points. As Ito and Okabe wrote, in 1990s Japan girls were associated with entering information technology through communication devices (pagers, mobile phones) and boys through games and computers. Generative AI entered workplaces less as a tool for talking with people than as a tool for instructing something to make things. This is closer to the side of the one who operates (①) and the computer side (③). One possible explanation is that men of this generation react strongly to AI because it is a revolution coming through the same entry point they used as children. No study, however, has tested this explanation directly.

Facts on the other side

Some facts keep the hypotheses from being stated strongly.

First, within the same generation there are people who push AI and people who do not use it. In Nakagomi and colleagues’ survey, the most common reason for not using generative AI was “I don’t feel the need” (39.9%), followed by “I don’t know how to use it” (18.5%). Among middle-aged and older people, “I don’t know how to use it,” “worries about security,” and “the environment for using it is not in place” were common (Chiba University press release). In MIC’s 2025 survey as well, 51.0% of people aged 40–49 and 55.8% of those aged 50–59 had never used it.

Second, the premise that generation determines views is itself criticized. Helsper and Eynon (2010), using a UK nationally representative survey, showed that generation is only one predictor of advanced internet use, and that breadth of use, experience, gender, and education are also important, in some cases more important than generation. The authors write that adults can become digital natives by acquiring skills and experience. Applied here, whether people in their 40s and 50s push AI may also depend less on birth year than on how many revolutions they have personally switched through (breadth of experience). If so, Hypotheses 1 and 2 fit the sources better when read as being about this breadth of experience rather than about birth year.

Third, the two surveys being compared differ. In MIC’s 2024 survey, the share with generative AI experience at ages 40–49 (29.6%) exceeded that at 30–39 (23.8%). In the 2025 survey, ages 30–39 (56.3%) exceeded 40–49 (49.0%). However, the survey design changed in 2025, for example by adding ages 15–19, so this reversal cannot be read as a change over the year.

Gaps

• Measuring advocacy: no study was found that measures “pushing” AI by age and gender. Counting adoption proposals, training initiatives, and statements in meetings by age would test Hypothesis 4.
• Japanese technology generations: Sackmann’s grouping rests on German diffusion curves. No study was found that redraws generations along Japanese diffusion timing.
• Transfer of the template: whether internet experience is used in evaluating AI could be tested by counting phrases like “back when the internet came” in narrative data by age.
• Gendered entry points: no study was found on whether men who entered through games and computers and women who entered through communication take different attitudes toward generative AI.

Related notes

• 90% Adoption x 10% Approval — The Paradox of AI Tool Diffusion and Evaluative Divergence: the structure in which the share of people who use AI tools and the share who view them positively diverge widely. It is the opposite problem to this note’s gap between use and advocacy.
• What Is the Frontier Model Premium Buying? (A Debate via Historical Analogy): a debate on paying for AI through historical analogies to technology. It can be read as one example of Hypothesis 2, using past futures in the present.
• Patterns of Technological Harm to Education, and Their Refutations: A Robustness Assessment Across 25 Works: the robustness of claims about harm from video games and smartphones. It corresponds to later tests of the adult discourse worrying about the young in ① and ②.

Unverified items

No main claim is unverified. For sources where only the abstract could be read (Humlum & Vestergaard 2025, Nakagomi et al. 2025, Kushida 2011, Vogel 2013, Jorgenson & Motohashi 2005), only what the abstract and press release say was used. The peak of pager subscriptions is read from a chart; no table of figures was available.

References

Statistics and official material

• Ministry of Internal Affairs and Communications (2026). White Paper on Information and Communications in Japan 2026, Summary (individual and corporate generative AI use, FY2025 survey). https://www.soumu.go.jp/main_content/001082851.pdf (main text https://www.soumu.go.jp/johotsusintokei/whitepaper/ja/r08/html/nd111310.html )
• Ministry of Internal Affairs and Communications (2025). White Paper on Information and Communications in Japan 2025, Summary (FY2024 survey). https://www.soumu.go.jp/main_content/001019264.pdf
• Ministry of Internal Affairs and Communications (2010). White Paper on Information and Communications in Japan 2010, Part 2, Chapter 4, Section 1 (internet users and penetration rate, Figure 4-1-1-1). https://www.soumu.go.jp/johotsusintokei/whitepaper/ja/h22/pdf/m4010000.pdf
• Ministry of Internal Affairs and Communications (2013). White Paper on Information and Communications in Japan 2013, Part 2, Chapter 4, Section 3 (household ownership of ICT devices, Figure 4-3-1-1). https://www.soumu.go.jp/johotsusintokei/whitepaper/ja/h25/pdf/n4300000.pdf
• Ministry of Internal Affairs and Communications (2025). Results of the 2024 Communications Usage Trend Survey (key points). https://www.soumu.go.jp/main_content/001011527.pdf
• Ministry of Internal Affairs and Communications (2019). White Paper on Information and Communications in Japan 2019, “The emergence and spread of the internet and changes in communication.” https://www.soumu.go.jp/johotsusintokei/whitepaper/ja/r01/html/nd111120.html
• Nintendo. Famicom timeline. https://www.nintendo.com/jp/famicom/history/index.html (accessed 2026-10-03)
• Shimanuki, Y. (1999). Paging Market Overview in Japan. NTT DoCoMo Technical Journal, 7(1), 6–8. https://www.docomo.ne.jp/binary/pdf/corporate/technology/rd/technical_journal/bn/vol7_1/vol7_1_006jp.pdf
• NTT DoCoMo. Overview of the i-mode service: An information distribution infrastructure for the 21st century (NTT DoCoMo Technical Journal Vol.7 No.2). https://www.docomo.ne.jp/corporate/technology/rd/technical_journal/bn/vol7_2/006.html (accessed 2026-10-03)

Children’s play and the social history of mobile phones

• Science Council of Japan, Subcommittee on Children’s Developmental Environments (2013). Recommendation: Toward improving children’s developmental environments in Japan: Issues and proposals on developmental time. https://www.scj.go.jp/ja/info/kohyo/pdf/kohyo-22-t169-3.pdf
• Tsuruyama, H., Hashizume, K., & Nakano, A. (2008). A study of the actual state of children’s play. Bulletin of the Faculty of International Studies, University of Toyama, 4, 133–137. https://www.tuins.ac.jp/common/docs/library/2008kokusai-PDF/0803tsuruyama2.pdf
• Ito, M., & Okabe, D. (2005). Mobile Phones, Japanese Youth, and the Re-placement of Social Contact. In R. Ling & P. E. Pedersen (Eds.), Mobile Communications: Re-negotiation of the Social Sphere (pp. 131–148). Springer. https://doi.org/10.1007/1-84628-248-9_9 (author version https://www.dourish.com/classes/ics234cw04/ito1.pdf )

Generations and expectations

• Sackmann, R., & Winkler, O. (2013). Technology generations revisited: The internet generation. Gerontechnology, 11(4), 493–503. https://doi.org/10.4017/gt.2013.11.4.002.00 (full text https://journal.gerontechnology.org/archives/1936-2079-1-PB.pdf )
• Krosnick, J. A., & Alwin, D. F. (1989). Aging and Susceptibility to Attitude Change. Journal of Personality and Social Psychology, 57(3), 416–425. https://doi.org/10.1037/0022-3514.57.3.416 (author version https://web.stanford.edu/dept/communication/faculty/krosnick/docs/1989/1989%20Aging%20and%20Att%20Change%20-%20Krosnick%20and%20Alwin.pdf )
• Helsper, E. J., & Eynon, R. (2010). Digital natives: where is the evidence? British Educational Research Journal, 36(3), 503–520. https://doi.org/10.1080/01411920902989227 (author version https://eprints.lse.ac.uk/27739/1/Digital_natives_%28LSERO%29.pdf )
• Brown, N., & Michael, M. (2003). A Sociology of Expectations: Retrospecting Prospects and Prospecting Retrospects. Technology Analysis & Strategic Management, 15(1), 3–18. https://doi.org/10.1080/0953732032000046024 (author version https://eprints-gro.gold.ac.uk/id/eprint/2379/2/SOC_Michael_2003a.pdf )

Generative AI use

• Bick, A., Blandin, A., & Deming, D. J. (2024, revised 2025). The Rapid Adoption of Generative AI. NBER Working Paper 32966 (peer-reviewed version not confirmed). https://doi.org/10.3386/w32966 (full text https://www.nber.org/papers/w32966.pdf )
• Humlum, A., & Vestergaard, E. (2025). The unequal adoption of ChatGPT exacerbates existing inequalities among workers. Proceedings of the National Academy of Sciences, 122(1), e2414972121. https://doi.org/10.1073/pnas.2414972121
• Nakagomi, A., Abe, N., & Tabuchi, T. (2025). Emerging generative AI divide: Personal, positional, and resource-based factors associated with use and reasons for non-use. Telematics and Informatics, 104, 102360. https://doi.org/10.1016/j.tele.2025.102360 (Chiba University press release https://www.chiba-u.jp/news/files/pdf/250113_AIdivide.pdf )

Japan’s ICT industry

• Kushida, K. E. (2011). Leading without Followers: How Politics and Market Dynamics Trapped Innovations in Japan’s Domestic “Galapagos” Telecommunications Sector. Journal of Industry, Competition and Trade, 11(3), 279–307. https://doi.org/10.1007/s10842-011-0104-7
• Vogel, S. K. (2013). What Ever Happened to Japanese Electronics?: A World Economy Perspective. The Asia-Pacific Journal, 11(45). https://doi.org/10.1017/S1557466013035043
• Jorgenson, D. W., & Motohashi, K. (2005). Information Technology and the Japanese Economy. Journal of the Japanese and International Economies, 19(4), 460–481 (NBER Working Paper 11801). https://www.nber.org/papers/w11801

Footnotes

1. Read from the bar chart; there is no table of figures. The same article describes the era when pagers were used only for emergency calls as one in which users felt “made to carry” pagers and “tied down” by them, and says that once pagers could display numbers and kana, the market expanded explosively to groups that had not used them before. ↩

2. The target age and estimation method differ between the figures up to the end of 2000 and those from the end of 2001. Up to the end of 1999 the target was ages 15–69, so these figures are not defined the same way as the figures from 2001, which cover ages 6 and over. ↩

3. Ito and Okabe also write that media reports linking kogyaru to pagers and compensated dating produced adult laments about young people’s moral decline. Pagers and mobile phones spread together with adult discourse worrying about the young. ↩