China is experiencing a convergence of population aging and digital transformation, with social aging and informatization becoming increasingly intertwined. Amid both digital dividends and digital divides, the health and social adaptation of older adults has emerged as a pressing public concerns (1, 2). According to the 55th Statistical Report on China’s Internet Development (2024), the number of Internet users aged 60 and above has reached 199 million, accounting for 64% of the elderly population (3). This finding suggests that, driven by continuous digital inclusion policies and the widespread adoption of information and communication technologies (ICT), the digital divide among older adults is gradually narrowing, creating new opportunities for social participation and personal development. Internet use among this group has increased markedly; their participation in the digital sphere has expanded; and online activities have become an integral part of their social interactions and daily lives (4–7). However, while bridging the digital divide and enhancing online engagement, frequent use of digital platforms may also foster emotional dependence and excessive immersion. Older adults are increasingly showing signs of digital addiction (8–10). Overuse of the Internet and digital platforms can disrupt sleep patterns, foster an “information cocoon” effect through compulsive short-video consumption, and lead to virtual interactions replacing real-world social contact (11–14). Such behaviors not only impair the psychological well-being and social functioning of older adults but also diminish their quality of life and hinder the realization of healthy aging goals. Consequently, digital addiction among older adults has become an emerging public health concern that requires immediate and concerted attention (15). To date, there is no consensus within the academic community on the definition of “digital addiction,” a term that is often used interchangeably with Internet addiction, smartphone addiction, and technology addiction (16). Goldberg (17) first introduced the concept of Internet addiction, defining it as a behavioral disorder in which excessive Internet use interferes with an individual’s daily life, studies, and work. Subsequent research has identified multiple characteristics of digital addiction, including adverse outcomes such as social isolation and neglect of real-world activities, as well as users’ immersive experiences and behavioral dependence (18–22). Some studies have even proposed time-based criteria, defining individuals who use digital devices for more than 4 h per day or over 30 h per week as addicted (23). From an aging research perspective, Du (24) defined digital addiction among older adults as a condition in which prolonged Internet use negatively affects their physical and mental health, as well as their social relationships. This study adopts the same definition. The existing literature has primarily examined the influencing factors of digital addiction among older adults from three dimensions: contextual, individual, and industrial. At the contextual level, the prevalence of digital information flows has reshaped social structures and symbolic environments, subtly transforming older adults’ values and lifestyles (25–27). At the individual level, many older adults have limited digital literacy, making it difficult for them to recognize manipulative design features embedded in digital platforms (28, 29). Driven by loneliness and emotional needs, they may voluntarily rely on digital devices and gradually become trapped in compulsive usage patterns (30, 31). At the industrial level, the rise of the “addictive digital economy” has encouraged enterprises to develop increasingly immersive and manipulative products aimed at capturing the “silver economy” (32–35), while insufficient regulatory oversight has further exacerbated the problem (36).

However, most studies on digital addiction have focused on its psychological factors (37), while overlooking the critical role of digital algorithms that now profoundly shape online behavior. Algorithmic recommendation systems and incentive mechanisms operate through both informational and psychological pathways, serving as important correlates of older adults’ excessive use of digital products (8, 38, 39). Nevertheless, the underlying mechanisms of these algorithms have not been adequately examined, and empirical evidence in this field remains scarce. To address this research gap, the present study constructs a behavioral model of digital addiction among older adults from an algorithmic perspective. Drawing on prior research on digital addiction, this study treats algorithmic inducement mechanisms (as perceived by participants) as the core independent variable. They are categorized into four dimensions: preferential incentives-profit-seeking psychological induction (40–42), interactive incentives-emotional compensation induction (8, 39, 43, 44), stage goals-feedback effect induction (45–47), and customized recommendations-exploratory psychological induction (48–50). The dependent variable is the level of digital addiction, while age, living arrangement, and daily smartphone use duration are included as control variables (51–54). Using a cross-sectional survey and regression analysis, this study empirically examines the associations between algorithmic incentives and digital addiction among older adults. It is hypothesized that all four algorithmic dimensions are positively associated with digital addiction in later life.

This study conducted a comprehensive analysis of the formation mechanism of digital addiction among Chinese adults aged 60 years and above, focusing on algorithmic inducement behaviors. The results, derived from a large sample size, revealed that digital addiction among older adults is not a sporadic issue but a widespread and structural one, with 77.11% of participants falling into the moderate or higher levels of addiction category. This finding should be interpreted with caution. Although the MPAI cutoffs proposed by Leung (2008) are widely used, validation studies specifically targeting older adult populations remain limited. In addition, given the sampling characteristics described above—particularly the reliance on convenience and purposive recruitment, and the high prevalence of intensive use of short-video platforms—the observed distribution may not reflect prevalence estimates in the broader older adult population. Nevertheless, this pattern suggests that digital immersion has become a salient phenomenon within this sample of older adults. More broadly, as Internet use becomes more prevalent, older adults are not only catching up in terms of technology adoption but also becoming increasingly dependent within an algorithm-mediated digital environment (61). The study found that all four dimensions of algorithmic inducement-preferential incentives, profit-seeking psychological induction, interactive incentives, emotional compensation induction, stage goals, feedback effect induction, and customized recommendations, exploratory psychological induction-were significantly and positively associated with digital addiction. Among these, interactive incentives–emotional compensation induction had the strongest effect, highlighting the critical influence of social feedback mechanisms in promoting media immersion among older adults (62). Preferential incentives–profit-seeking psychological induction ranked second, indicating the continued importance of economic stimulation in influencing usage behaviors in this group (63). Although stage goals–feedback effect induction and customized recommendations–exploratory psychological induction showed relatively weaker effects, both remained statistically significant. The model’s coefficient of determination (R² = 0.791) indicated that the four dimensions, together with the control variables, explained 79.1% of the variance in digital addiction, demonstrating a good overall model fit. Although the model accounts for a substantial proportion of variance (R² = 0.791), the findings should be interpreted with appropriate caution. Because all variables were measured using self-report instruments at a single time point, some degree of shared-method variance cannot be entirely ruled out, and the cross-sectional design does not permit firm causal conclusions. A Harman single-factor test indicated that the first unrotated factor accounted for 52.51% of the variance, suggesting that common-method variance may be present. Nevertheless, the consistency of the results across multiple algorithmic dimensions provides convergent support for the robustness of the associations observed.

The findings indicate that digital addiction among older adults is not merely a psychological phenomenon but rather the result of multiple structural factors acting in combination, specifically reflecting the interplay of individual, technological, and institutional dilemmas (64–66). Older adults, constrained by cognitive decline and weakened social support, are more likely to develop psychological dependence within algorithm-mediated digital environments (64, 67, 68). Meanwhile, algorithms operate under the core logic of “attention maximization” (69), reinforcing usage behaviors through reward feedback and personalized recommendations (70–72). Immediate reward engagement and time preferences play significant roles in compulsive use behaviors. Hyperbolic discounting accounts for this phenomenon by positing that individuals tend to prioritize short-term rewards over long-term gains (73). While existing research suggests a correlation between these mechanisms and addictive behaviors (74, 75), it remains to be further validated whether they directly lead to compulsive use. Digital platforms, particularly recommendation systems, exploit this by offering continuous feedback, triggering immediate gratification, and reinforcing repetitive behaviors. Manipulative platform designs further intensify this by enhancing user engagement through constant feedback mechanisms, solidifying the cycle of addiction (76). This process involves reward learning and cue reactivity, where users become increasingly dependent on instant rewards, leading to habitual and compulsive use (77). At the institutional level, the existing digital governance system remains insufficient in terms of algorithmic ethics and user protection. Weak regulatory mechanisms and low algorithmic transparency expose older adults to risks of commercial manipulation and content control (78, 79).

Specific mechanisms, such as preferential incentives–profit-seeking psychological induction, showed a significant positive association with digital addiction among older adults (B = 0.190, p = 0.001). The data revealed that 54.23% of respondents reported that they “purchase as soon as possible when seeing notifications of limited-time offers or product scarcity,” and 59.13% believed that “flash sales or purchase restrictions make them shop more frequently.” These findings highlight the significant influence of economic incentives on the behavior of older adults. Digital platforms, through coupons, limited-time discounts, and point-based reward tasks, effectively create a sense of “scarcity,” triggering immediate behavioral responses (80–82) and reflecting the sophisticated strategies employed in digital marketing. At the psychological level, older adults have long been influenced by frugality-oriented values such as “thriftiness” and “saving whenever possible.” When facing algorithmically constructed situations of virtual scarcity, they are more likely to experience “opportunity anxiety,” leading to impulsive purchasing and repetitive use behaviors (83–85). The “scarcity effect” (86, 87) and “loss aversion theory” (88) in behavioral economics provide a sound theoretical explanation for this phenomenon. As algorithmic mechanisms continuously amplify these economic stimuli, rational saving behavior gradually transforms into irrational immersion, forming an economically driven pathway to digital addiction and illustrating the potential risks associated with this phenomenon.

Interactive incentives–emotional compensation induction, referring to the emotional rewards or gratifications that individuals receive from digital interactions, had the most pronounced effect (B = 0.293, p = 0.001), serving as a core variable most strongly associated with digital addiction among older adults. The survey results indicated that 59.13% of respondents agreed that “likes, comments, and other feedback encourage repeated use of the application,” while 59.67% reported that “receiving replies from others or gaining followers brings a sense of pleasure.” These findings suggest that algorithm-enabled platform features may reinforce older adults’ media engagement through emotional incentive mechanisms. Due to the weakening of social roles and limited real-world interactions, older adults are more inclined to seek emotional feedback and a sense of social presence through online communication (89–91). Features embedded in algorithms, such as likes, comments, and follower notifications, provide immediate feedback and psychological gratification, partially compensating for diminished social roles (92, 93). However, the immediacy and selectivity of such feedback may make older adults more susceptible to virtual dependence, which may gradually evolve into emotionally driven digital addiction (55).

Third, stage goals–feedback effect induction was also significantly and positively associated with digital addiction among older adults (B = 0.142, p = 0.001). A total of 61.04% of respondents indicated that they “try to complete platform tasks to avoid the expiration of rewards or coins,” while 62.94% reported that they “often or always continue watching short videos to learn about subsequent content.” This mechanism can be explained by the Zeigarnik effect in psychology, which posits that unfinished tasks create psychological tension and a persistent drive, motivating individuals to repeat behaviors until a sense of completion and psychological equilibrium is achieved (94). Digital platforms transform this tension into motivation for continued engagement through mechanisms such as point rewards, task unlocking, and level systems, leading older adults into a behavioral cycle of “unfinished tasks–repeated logins–reinforced feedback” (95–97). This task-driven mechanism substantially increases usage frequency and, over time, strengthens media dependence and habitual behavioral patterns (98).

Finally, customized recommendations–exploratory psychological induction also showed a significant effect (B = 0.149, p = 0.001). A total of 64.31% of older adults reported that they “continue browsing because they anticipate the next video,” and 65.67% agreed that “recommended content aligns with their personal interests and needs.” Algorithmic recommendation systems, driven by big data analytics and user profiling, continuously optimize their delivery models to achieve precise content matching (99, 100). While this mechanism reduces information search costs, it simultaneously narrows the diversity of information sources (101). Curiosity and the desire for recognition further drive older adults to become immersed in content environments shaped by algorithmic recommendations (64, 102). As recommendation systems self-optimize, platforms establish a closed loop of “interest modeling–content delivery–behavioral feedback–algorithmic reinforcement,” gradually leading users to develop personalized dependence (99). The dual nature of customized recommendations lies in their ability to enhance user experience while simultaneously reinforcing the “information cocoon effect,” fostering a continuous exploratory form of addiction at the psychological level (13, 103). Overall, algorithmic inducement behaviors, through four psychological pathways-economic stimulation, which involves the anticipation of rewards; emotional compensation, which provides a sense of comfort or satisfaction; goal-driven feedback, which reinforces the achievement of specific objectives; and exploratory reinforcement, which encourages the discovery of new content-are jointly associated with a shift in older adults’ digital usage from instrumental participation to emotional dependence (104). The intertwining of psychological compensation and technological reinforcement constitutes the core mechanism underlying the formation of digital addiction among older adults.

Furthermore, this study provides important implications at both theoretical and practical levels. Theoretically, it proposes an analytical framework of “algorithmic drive–psychological compensation–structural constraint–behavioral reinforcement,” revealing the underlying logic of digital addiction. This framework is intended as an interpretive lens for organizing the observed associations rather than a causal model. Digital addiction among older adults is not merely a technological outcome but the result of interactions among psychological needs, algorithmic mechanisms, and social structures. This finding expands the generational perspective of digital addiction research and offers a new theoretical foundation for understanding the psychological vulnerability and media adaptation mechanisms of older adults in an algorithmic society. In particular, understanding the interplay between algorithmic designs and older adults’ psychological traits is critical in addressing algorithmic exploitation and addiction risks (105). Older adults, often less equipped to recognize persuasive digital cues, face heightened risks of algorithmic manipulation, which may exacerbate their vulnerability to digital addiction (7, 39). In practice, the findings have significant implications for governments, platforms, and social governance. Governments should strengthen regulatory frameworks for algorithmic ethics and user protection to prevent commercial platforms from exploiting addictive design features that may harm vulnerable groups (66, 106). More specifically, regulatory agencies could develop age-sensitive guidelines limiting the intensity and frequency of reward-based inducements, require clearer disclosures about how recommendation algorithms prioritize content, and establish auditing mechanisms to evaluate potential risks to older adults. Platforms should improve algorithmic transparency, reduce reward-based inducements and excessive personalized recommendations, and implement reminders that promote healthy usage habits (107, 108). For example, platforms could offer optional low-stimulation modes for older adults, provide gentle prompts during prolonged screen time, and design simplified controls allowing users to adjust or disable personalized recommendations. Communities and families should enhance digital literacy education and provide psychological support to help older adults identify algorithmic inducement behaviors, fostering rational internet use and balanced social engagement (109, 110). Concrete literacy training could include teaching older adults how to recognize persuasive cues (such as scarcity prompts or emotional feedback), manage notifications, and set personal usage boundaries. In conclusion, coordinated governance across multiple sectors can achieve a balance between technological advancement and social protection, thereby fostering an inclusive and safe digital ecosystem that promotes active aging and advances the goal of digital well-being.

This study examined the formation mechanism of digital addiction among Chinese adults aged 60 years and above from the perspective of algorithmic inducement behaviors. The findings revealed a dual pattern in the digitalization process of older adults-bridging the digital divide while falling into digital dependency. Platform design features and algorithm-enabled mechanisms, through mechanisms such as economic incentives, social feedback, task incentives, and content recommendations, may fulfill the psychological compensation needs of older adults while simultaneously reinforcing habitual use and emotional attachment. These mechanisms collectively may help explain the transition from instrumental to immersive digital engagement. Overall, the study uncovers the generative logic of digital addiction among older adults: the external inducement of algorithmic drivers interacts with individual psychological compensation needs, forming a cyclical mechanism of behavioral dependence and immersive use under the structural constraints of technological logic and institutional environments. These findings suggest that digital addiction among older adults is not merely a psychological or technological issue but rather the outcome of the combined effects of algorithmic environments, social support, and governance mechanisms. Theoretically, this study extends the generational perspective in digital addiction research and proposes an explanatory framework- “algorithmic drive–psychological compensation–structural constraint–behavioral reinforcement.” This framework provides a new analytical lens for understanding media adaptation and psychological vulnerability among older adults in algorithmic societies. In practice, the study emphasizes the importance of collaboration between digital governance and social support. A balance should be sought between technological innovation and social protection to prevent algorithmic inducements from intensifying digital dependence among older adults and to promote the development of an inclusive, safe, and sustainable digital ecosystem for the aging population. Despite revealing the association between algorithmic mechanisms and digital addiction through survey data, this study has certain limitations. First, although the study identifies strong associations between algorithmic inducement and digital addiction, its cross-sectional design does not allow for firm causal inferences. Second, although both online and offline recruitment were used, the sampling strategy may still overrepresent older adults who are more digitally active or more familiar with smartphone-based surveys, thereby limiting representation of less-connected groups. Third, all variables relied on self-report measures rather than behavioral log data; thus, recall bias and shared-method variance cannot be entirely ruled out. Fourth, the study was conducted within the sociocultural and platform ecosystem of mainland China, and the dynamics of algorithmic inducement may differ in other cultural or regulatory contexts. Future research could employ longitudinal tracking and behavioral log data to further explore the dynamic relationships between algorithmic feedback and psychological dependence. It should also focus on the long-term impacts of digital environments on older adults’ cognition, social interactions, and quality of life, thereby providing stronger empirical support for healthy aging and digital governance in an increasingly algorithmic society.