Age and Domain Differences in Attitudes Toward Artificial Intelligence

Artificial intelligence (AI) is moving quickly into everyday life, from workplace automation to healthcare and education. Public conversations often frame people as either enthusiastic or worried about AI, but our research shows attitudes are more complicated than that.

We surveyed one hundred participants and measured their views using the ATTARI-WHE framework (Work, Healthcare, Education), which was recently validated by Yeo and colleagues⁸. This framework separates attitudes into three parts: cognitive (what people believe about AI’s usefulness), affective (how people feel about AI), and behavioral (whether people are willing to use AI). By looking at all three together, we can see not just what people think, but how that translates into action.

Our results echo what Yeo et al. reported. People generally believed AI was useful (high cognitive scores), but they were much less likely to say they would use it (low behavioral scores). This gap between belief and behavior shows that simply knowing AI has value is not enough to make people adopt it. Education was the most positively rated domain, especially in terms of beliefs. Work-related AI was rated more moderately as applicable but not embraced in practice. Healthcare was the most polarized: people in their fifties were positive both in their beliefs and feelings, while those in their forties stood out as particularly skeptical.

The clearest new finding from our study was the role of age. People in their forties consistently gave the lowest ratings across domains and facets, while people in their thirties and sixties were more positive, especially in education. Those in their fifties were unusual in showing especially positive views of healthcare AI. These patterns suggest that age shapes how people see AI in ways that earlier validation work did not capture. We also analyzed other demographic factors, including gender, education, employment, income, and ethnicity, and found no significant effects, reinforcing that age was the key factor.

Why might people in their forties be more skeptical?

Skepticism in this group matches what other studies have found about midlife attitudes toward technology and well-being. Midlife often brings the most career and family responsibilities, which can make people more cautious about disruptive change. Research has shown that workers in mid-career are particularly sensitive to automation because they have much to lose yet many years left in the workforce⁴˒⁷. Studies on risk perception also suggest that heavier responsibility loads make people more cautious about new technologies⁶.

Sociologists describe today’s forty-somethings as a “digital transition generation.” Unlike younger adults who grew up with digital technologies or older adults who can opt out, this group has repeatedly had to adapt to waves of technological change. This can create “technology fatigue” and skepticism⁵. In addition, research on well-being consistently shows a dip in satisfaction during midlife, which may heighten skepticism about new and uncertain changes¹˒²˒³. These overlapping pressures provide a compelling explanation for why the forties cohort in our study showed the most negative attitudes toward AI.

Implications

Our findings carry two main implications. First, the belief–behavior gap means that showcasing AI’s capabilities will not be enough to drive adoption. Building trust, offering transparency, and demonstrating clear benefits are essential. Second, because attitudes vary by age and domain, strategies for introducing AI need to be tailored. Education is likely to be a welcoming entry point, while healthcare will require more careful engagement, particularly for midlife adults.

By applying the ATTARI-WHE framework to a new dataset, we show that attitudes toward AI are not uniform. They depend on the domain, the type of evaluation (belief, feeling, or action), and demographic context–most notably age. Recognizing these differences is essential for policymakers, educators, and healthcare leaders who want to bridge the gap between seeing AI’s value and using it in real-world settings.

---

References

1. Blanchflower, D. G., & Graham, C. (2020). The Midlife Dip in Well-being: Evidence from Multiple Countries. Journal of Economic Behavior & Organization, 176, 317–335.

2. Blanchflower, D. G. (2022). Is happiness U-shaped everywhere? Age and subjective well-being in 145 countries. Journal of Population Economics, 35(2), 529–561.

3. Brookings Institution. (2020). The midlife dip in well-being: Why it matters at times of crisis. https://www.brookings.edu/articles/the-midlife-dip-in-well-being-why-it-matters-at-times-of-crisis/

4. Frey, C. B., & Osborne, M. A. (2017). The Future of Employment: How Susceptible Are Jobs to Computerization? Technological Forecasting and Social Change, 114, 254–280.

5. Hargittai, E. (2010). Digital Na(t)ives? Variation in Internet skills and uses among members of the “Net Generation.” Sociological Inquiry, 80(1), 92–113.

6. Slovic, P. (1987). Perception of risk. Science, 236(4799), 280–285.

7. Susskind, R., & Susskind, D. (2015). The Future of the Professions: How Technology Will Transform the Work of Human Experts. Oxford University Press.

8. Yeo, M., Kim, J., & Choi, S. (2024). Development and validation of the ATTARI-WHE: A multidimensional scale for measuring attitudes toward artificial intelligence in work, healthcare, and education. Patterns, 5(5), 100918. https://doi.org/10.1016/j.patter.2024.100918