A Grand Challenge for Psychology: Reducing the Age-related Digital Divide

Abstract

World-wide population aging and rapid diffusion of digital technology have converged to produce an age-related digital divide in technology adoption, as seen in use of the internet and ownership of smartphones. Given the centrality of these technologies for full participation in modern society, reducing that gap is an important challenge for psychologists. We outline more and less malleable factors associated with technology adoption. We argue that interventions that can change both the aging user and the design of products will be necessary. Adaptive technology systems that incorporate artificial intelligence and extended reality represent promising new approaches to reducing the age-related digital divide.

Demographers project that worldwide the older adult population (age 65+) will grow from about a billion older adults in 2030 to 1.6 billion by 2090, and the 65+ age category is projected to double from nine to 18% of the world’s population between 2020 and 2100 (https://population.un.org/wpp/Download/Standard/Population/, accessed 3/7/2021).

Rapid population aging has been accompanied by a second trend: the acceleration of digital technology diffusion (Charness, 2017). Their convergence has led to a persistent age-related digital divide, first recognized in 1995 (https://www.ntia.doc.gov/ntiahome/fallingthru.html) and still seen 25 years later (https://www.ntia.doc.gov/blog/2020/ntia-data-reveal-shifts-technology-use-persistent-digital-divide). A digital divide is usually defined as a difference in digital technology adoption and use as a function of person characteristics such as age (e.g., younger/older), race, income level, and location (e.g., rural vs urban dwelling). Examples of this age gap can be seen for internet access and smartphone ownership in Figure 1, showing population-representative data for the USA from Pew Research Center (2021a, 2021b).

Figure 1.

Top panel depicts the percentage of people in each age group (18–29, 30–49, 50–64, 65+) who reported using the internet, with data collected from the year 2000 to 2021. Bottom panel depicts the percentage of people in each age group (18–29, 30–49, 50–64, 65+) who reported owning a smartphone. Data from Pew Research (2021a, 2021b).

A similar technology adoption gap exists in many other countries (e.g., in Japan, the country with the world’s oldest population (Figure 5-2-1-4 in https://www.soumu.go.jp/johotsusintokei/whitepaper/eng/WP2020/chapter-5.pdf#page=11). Those without internet access and digital skills can become seriously disadvantaged. Many United States SARS-CoV-2 vaccine distributors have relied on automated web sites for scheduling appointments rather than on telephone-based access that requires more staff. More than a quarter of adults age 65+ years in the United States were not on the internet in 2019, and as seen in Figure 1, between 2019 and 2021 internet use among older adults rose only 2% (Pew, 2021a).

Although we focus on aging adults in this review, we note that digital divides are also present for minority groups compared to majority ones, less educated and poorer people compared to more educated and wealthier ones, and rural compared to urban dwellers (Pew, 2021a). Intersectionality is an important issue to consider. Aging adults also falling in those other disadvantaged categories are at even greater risk of being unable to cope with digital challenges such as those posed by the recent pandemic (Xie, Charness, Fingerman, Kaye, Kim…, 2020). It is also important to acknowledge that predictive factors such as cohort, gender, and experience with technology systems are changing over time, for instance, in predicting attitudes toward computers (Lee et al., 2019).

Earlier, we suggested (Charness & Boot, 2009) that the digital divide for aging adults is unlikely to close on its own, partly because of a mix of modifiable and difficult to modify factors: modifiable technology design features, and person features some of which are more modifiable, for example negative attitudes, and others that are less easily modified such as age-related changes in cognitive, perceptual, and psychomotor abilities. We concentrate here on recent progress in surmounting such barriers. As Figure 2 illustrates, we believe that psychologists have an important role to play in meeting the challenge of reducing this digital divide. Reducing the gap requires a detailed understanding of the complex and interconnected attitudinal, perceptual, cognitive, and experiential factors that contribute to it.

Figure 2.

Factors related to technology adoption and use and fields of psychology that can contribute to the grand challenge of reducing the age-related digital divide. Advocacy by psychologists can also shape factors such as education and affordability.

We next discuss person factors influencing adoption and use.

Attitudinal Barriers.

General attitudes continue to contribute to a lack of use and adoption of existing and emerging technologies by older adults, including attitudes such as technology self-efficacy, comfort, and interest. An analysis of data collected from 1994 to 2013 found large differences between younger and older adults on these attitudes that can help account for the digital divide, and unfortunately, little evidence that individual older adults’ attitudes changed over this nearly two decade long period ending in 2013 (Lee et al., 2019). Instead, there was evidence for younger adults aging into older age categories and bringing their more positive attitudes with them. Although attitudes are modifiable, they may not change spontaneously over time, highlighting the importance of psychological interventions that target attitudes related to specific technologies. One approach might be highlighting potential benefits, thereby influencing perceived usefulness. Highlighting attitudes as a target that continues to be relevant, age-associated attitude differences have been observed related to even newer technologies, including autonomous vehicles (Rovira et al., 2019). Familiarity may play an important role in shaping attitudes, more related to cohort differences than age per se; younger adults have been exposed to a multitude of digital technologies at an early age, in the home, at school, and in the workplace while many older adults today, especially those among the “oldest old” (age 85+) lacked these experiences during development.

Cognitive Barriers.

Given the complexity of many new technologies, and that many of these technologies involve small screens and devices, even normative changes in cognition and perception (e.g., changes in visual acuity and working memory), can negatively influence older adults’ interactions with technologies and technology attitudes, and these issues can be exacerbated by age-related disease processes that impair cognition (Schmidt & Wahl, 2019). Adoption can also be impacted by age-related differences in the cost of new learning; age influences the rate of learning of new technologies, especially technologies very different from technologies one already has experience with, meaning that older adults must invest more time and effort when learning new technology compared to younger adults, which serves as a barrier (Sharit, Moxley & Czaja, 2021).

Design Barriers.

Technology design that does not consider the needs, preferences, and abilities of older adults can play an important role, affecting older adults’ perceived usefulness and usability of technology. In the design of technology, lack of consideration of older adults may stem from the demographics of those creating technologies. In the United States, for example, most software developers are in their 30s and 40s, much younger than the general employed population (Figure 3). Given egocentric bias in reasoning about others’ knowledge and abilities (Nickerson, 1999), we can expect that younger designers would have trouble imagining product use from the perspective of an aging adult.

Figure 3.

Age distribution (percent) of software developers, all employed persons. Data from the 2020 population survey of the U.S. Bureau of Labor Statistics, https://www.bls.gov/cps/cpsaat11b.htm accessed 3/29/2021.

Frameworks for Understanding Technology Use and Adoption.

Most models of technology adoption and use emphasize that a potential user weighs perceived benefits and costs, factors with positive and negative valences (e.g., Unified Theory of Acceptance and Use of Technology; Venkatesh, Thong, & Xu, 2012). Since the publication of our original article, models have been developed that specifically consider age-related factors that impact technology adoption (e.g., Senior Technology Adoption Model, Chen & Lau, 2020). Thinking of technology adoption and use models, we can envision designing interventions for common components such as perceived benefit and perceived cost, for instance changing advertising materials to make benefits more salient. Other specific factors identified in these and other models include social influence (extent to which friends and family use technology), facilitating conditions (availability of technology support), hedonic motivation (potential fun or pleasure using technology), price value, habit (cost of changing from old to new methods), and trust.

Interventions need to address these components. For younger age cohorts, internet connectivity is near 100%, implying that benefits (internet-accessible resources) generally outweigh costs (price of connectivity, privacy risk). Data collected about non-internet users (Zickuhr, 2013) show that the top categories cited for nonuse were perceived lack of relevance, poor usability, high price, and lack of access/availability. In the field of human factors (Salvendy, 2012), interventions focus on changing the device or system through improved design, changing the user through training, or both. Assuming nonusers were truthful in their survey responses, it appears necessary to adopt a multi-pronged approach to address barriers leading to nonuse, for instance providing better training options, improving design, and subsidizing costs.

One user constraint, inability to see the relevance of technology, is likely to be a diminishing problem. An example is the recent imperative to access many goods and services, including healthcare visits that became telehealth visits, during the Sars-CoV-2 pandemic. Design guidelines to improve usability of technology for older adults are widely available (e.g., Czaja, Boot, Charness & Rogers, 2019). Lowering monetary costs for technology adoption would involve policy changes such as subsidizing connectivity to the internet both in terms of access (e.g., rural vs urban) and cost for technology devices and services. There is precedent for national policy changes to improve infrastructure and accessibility in the United States, initially for national electrification and later for telephone ownership.

Some predictors of technology adoption present major challenges. Older adults have widely varying income, education, and technology experience levels. Further, closing the digital divide at the level of the individual requires some understanding of their current proficiency (e.g., Roque & Boot, 2018, 2020) to tailor interventions. Among modifiable factors, improving technology proficiency levels with well-designed training is both viable (Czaja, Boot, Charness, Rogers & Sharit, 2017) and desirable for improving technology self-efficacy. There is evidence that training is more successful when it is procedural, providing step-by-step instructions rather than conceptual training that attempts to provide schemas for understanding how an application works (Hickman, Rogers & Fisk, 2007). Another individual difference variable, cognitive ability, is a strong predictor of breadth and depth of technology use (Czaja, Charness, Fisk, Hertzog, Nair… et al., 2006). However, interventions to improve cognition in older adults have had little impact in terms of transfer to non-trained tasks (Simons, Boot, Charness, Gathercole, Chabris…et al., 2016), underlining the need to focus on training specific technology skills.

Promising Approaches.

Fortunately, we are on the cusp of the greater diffusion of emerging technologies that have the potential both to help close the age-related digital divide and provide substantial support to benefit older adults’ health, wellbeing, and independence, including innovations in artificial intelligence (AI) and extended reality (XR). At the start of the microcomputer revolution, when user interfaces relied mainly on recalling and typing in commands, Carroll and Carrithers (1984) argued for a “training wheels” approach of simplifying the interface to prevent user errors, and as a user gained experience, revealing features that permit more actions.

A promising approach today is to design technology products that adapt to the experience and capabilities of the user, making use of machine-learning techniques to provide “just-in-time” support in the form of prompts or advice when difficulties are likely to be encountered. This approach may be especially beneficial for older adults with time-varying needs for support due to age-associated cognitive impairment or dementia. Based on the user’s history of interactions with a system (and individuals like them), the system might anticipate the intention of the individual and help guide the user through the necessary steps to carry out that intention when the system detects they are having trouble.

Another promising area of investigation is the potential of AI to help support adherence to various health behaviors, including medication management, physical exercise, and home-based health and cognitive assessments. Traditional technology-based adherence support systems have demonstrated limited efficacy at boosting adherence (Mistry et al., 2015). However, intelligent and adaptive systems can be used to remotely monitor intervention engagement and provide customized adherence support based on the history, characteristics, and motivations of the user, providing the right message delivered at the right time to be maximally effective (Charness, Boot, & Gray, 2020).

The potential utility of AI-based systems is shown with today’s voice-based digital assistants. Older adults, when trained to interact with them, tend to like home smart speaker systems, despite their inability to respond in a useful way to about a third of queries (Kim & Choudhury, 2021). However, such devices require a complex infrastructure, smartphones for setup and home Wi-Fi networks, and aging adults today report low proficiency in managing both those technologies (e.g., Roque & Boot, 2016; 2020).

Extended reality (XR) systems, which encompasses virtual reality (VR) and augmented reality (AR), are now advanced enough and inexpensive enough for these technologies to be explored as methods to help support older adults in their everyday lives. AR in particular has the potential to overlay instructions and guidance on the view of technology-based tasks to assist and support the use of technology, as well as the performance of non-technology-based tasks. VR has potential to train the use of emerging technologies such as autonomous vehicles in a risk-free environment (e.g., Sportillo, Paljic, & Ojeda, 2018). In addition to facilitating the learning of new technologies to reduce the digital divide, XR may have the potential to support several activities to enhance well-being and promote mental and physical health. However, for both AI and XR solutions to be effective, these technologies themselves need to be studied to understand barriers to use and adoption by older adults, and these are two important and emerging fields of study.

The Future.

Even with unlimited resources and the best intervention science to try to counter the age-related digital divide, it is questionable whether that divide will ever vanish. Person factors such as an age-related learning rate decline (memory system degradation) coupled with period effects such as technology diffusion acceleration, factors that tend to widen the digital divide, seem poised to maintain this digital divide as a challenge for psychological science for at least the next few decades. However, psychological scientists have well-developed tools, methods, and models that will serve invaluable roles now and in the future in addressing technology design challenges. They can support individuals of all ages to obtain the benefits of existing and emerging technologies, possibly reducing the digital divide and enhancing digital equity in an aging world (https://www.un.org/development/desa/ageing/2021-unidop-digital-equity-for-all-ages.html#_ftn1).