AI is increasingly used in consumer wellness technologies for health-related purposes [1]. Digital platforms can collect continuous behavioral and physiological data, infer lifestyle patterns, and deliver personalized recommendations [2]. Advances in hardware and generative AI enable real-time feedback loops that can adapt nutrition and physical activity guidance to individual goals, preferences, and contexts. As a result, AI-powered wellness apps are becoming persistent, interactive agents in users’ daily lives rather than passive self-monitoring tools. The widespread ownership of smart devices enables a rapid expansion of AI-enabled diet and exercise apps. A recent study estimates that more than 600 million users own wearables capable of tracking their heart rate and step counts, and 300 to 500 million users own a sleep-tracking device [3]. Consumer demand creates opportunities for scalable behavior change support for weight management, physical activity promotion, and chronic disease prevention [4].

As consumer wellness technologies increasingly influence prevention, behavior change, and chronic disease risk management, they represent an important frontier for medical futures research. These technologies are often framed as a form of preventive health intervention that may reduce reliance on clinical care and mitigate growing health care costs. At the same time, their widespread adoption introduces complex challenges related to data privacy, algorithmic bias, user agency, and unequal access to digital resources [5,6]. As AI systems increasingly shape everyday health behaviors outside traditional medical settings, understanding their broader and longer-term consequences for users has become an urgent research priority. This demand aligns with public health priorities supporting healthy longevity in an aging society.

In the context of this rapidly changing landscape with well-identified actors and emerging technologies, there is a need for a structured overview of possible consequences and implications. Futures studies, an interdisciplinary discipline that uses different foresight methodologies to explore and anticipate future developments, provide a forward-looking framework for exploring the future of diet and fitness apps powered by AI [7]. Futures methodologies have been used for decades in many fields, but they remain underutilized in the health care domain [8]. Given the accelerated technological progress in health care and the subversion of health care roles brought about by digital health, a systematic study of the future of wellness apps is warranted. This paper aims to investigate the future of AI in diet and exercise apps using a foresight methodology known as the Futures Wheel (FW)—a strategic foresight tool that visually maps the direct and indirect consequences of a central trend—by identifying first- and second-order ripple effects that form the consequences of AI integration into consumer health apps [9,10]. The FW is particularly suited to the analysis of a user-facing digital health technology domain where complex interactions generate nonlinear effects and several potential scenarios.

The FW proved well suited to the analysis of consumer-facing AI wellness technologies: its consequence-mapping structure accommodated the nonlinear and multistakeholder nature of AI wellness technology diffusion, while the 2-tier causal architecture maintained traceability between the central trend and downstream effects. The FW is well positioned to reveal structural interdependencies across consequences rather than converging on consensus predictions—an advantage in an area characterized by high uncertainty and rapidly shifting technological baselines.

Our FW shows that integration of AI into wellness apps may generate cascading effects across behavioral, technological, institutional, and ethical domains. These effects do not point to a univocal future but to multiple interacting trajectories, including improvements in health outcomes, surveillance risks, institutional adaptation, and digital inequality.

At a user level, the most immediate first-order consequences identified by the wheel are the individualization of nutrition and fitness plans and constant health coaching, made possible by AI’s increasing data analysis and adaptive recommendation capabilities. For the same reason, improvements in healthy behavior adherence and preventive health scalability are expectable. Passive tracking tools become agents that identify behavioral triggers or early signs of health issues. Second-order consequences, however, warn of the psychological risks of AI wellness implementation. Technostress, dependency, and reduced intrinsic motivation may be the trade-offs of an otherwise health-promoting platform.

The wheel highlights technological convergence as another enabler of future developments. The integration of wearables, smart environments, and AI analytics expands wellness into new environments with ambient intelligence.

Demands for data protection are likely to increase, as AI platforms continuously collect sensitive data. Monetization, surveillance, or user profiling may be another trade-off of AI expansion into wellness. In parallel, protective technical and legal frameworks will be developed, but the tension will persist. In the same vein, the FW identified challenges in data-driven risk profiling and institutional use of behavioral data. Insurance underwriting and corporate wellness programs may incentivize healthy behaviors, but there is a risk of discrimination against those with socioeconomic constraints or disabilities, who might face higher premiums or workplace penalties. As illustrated in the domain analysis, the risks identified in this branch of the wheel are unlikely to be meaningfully constrained until the regulatory and ethical domain reaches at least an intermediate implementation level—suggesting that proactive regulatory action, rather than technological self-correction, is the more plausible near-term safeguard.

AI platforms meant to promote health may intensify health inequality. Mobile device ownership, platform subscription, and digital skills may contribute to the digital divide within the population. Equalizing initiatives also emerge on the wheel, such as subsidized devices or inclusive design.

AI wellness technologies may change practices related to food and exercise. Individual diets that ignore social context complicate communal eating, and individual routines may clash with group activities. Health optimization may take precedence over social enjoyment, altering cultural perceptions of eating or exercising.

The FW suggests that the expansion of AI wellness may be shaped by institutional and regulatory responses. The frontier between consumer tools and digital therapeutics may become increasingly blurred, prompting institutions to build frameworks for safety, transparency, and accountability. Institutional integration brings scale, sustainability, and trust, and will play a role in legitimizing AI presence in population health strategies. Conversely, institutional interests may conflict with particular ones, raising concerns about matters such as consent, privacy, and surveillance.

AI demand for computational power is increasing, prompting concerns around carbon dioxide emissions, water use, and rare mineral extraction. However, quantifications vary [15-18]. To some extent, these developments could diminish the positive implications of AI wellness. Because of its uncertain and cross-cutting nature, this effect was not included in any FW branch.

Users are the focus of this study, but other stakeholders will also play an important role in shaping and supporting the implementation of AI wellness solutions. The consequences outlined by the FW will require that developers increasingly include users in the design phase, mitigate algorithmic bias, and preempt negative behavioral effects. Regulators will need to create adaptive frameworks that do not stifle creativity but address safety and indication creep. Health care providers will have to appraise the integration of AI apps into clinical workflows, especially regarding logistics and data reliability.

In the future, improvements will necessarily happen in algorithm accuracy for diverse users. To enhance user experience and weight loss outcomes, AI chatbots will probably be designed to be human-like and context-aware. Apps will integrate health data, user traits, behavioral patterns, and emotional states to deliver effective, tailored recommendations. As predictive models grow more precise, apps may begin delivering real-time interventions based on subtle behavioral cues or environmental data. On the other hand, AI is expected to become more adaptive and persuasive and might blur the line between support and manipulation. User agency and informed consent might inform the future ethical design of feedback systems. International policy alignment through the Organisation for Economic Co-operation and Development (OECD) AI frameworks or the European Union Health Data Space may materialize in equitable and safe AI development.

Taken together, the first- and second-order effects of this FW indicate that we might be approaching a sociotechnical transformation rather than a technological shift. Apart from users, the consequences affect institutional structures, data governance, and cultural norms. A thoughtful implementation should maximize the potential of AI integration in lifestyle routines while ensuring ethical and sustainable use.

Rigor in the foresight process was aided by multiple safeguards. First, the panel members independently generated first-order consequences before the initial group discussion to reduce potential anchoring and promote diverse viewpoints. Second, the FW followed a predefined 2-tier causal structure linking first-order and second-order consequences to the central trend, ensuring traceability. Third, consequences were included only after unanimous agreement. This principle provides a plausibility filter intended to exclude effects considered insufficiently justified or speculative. The interdisciplinary composition of the panel enabled consideration of physiological, behavioral, and sociotechnical dimensions of AI-enabled wellness technologies.

Nevertheless, certain limitations should be acknowledged. First, the panel consisted of 3 experts who were also the study authors, which may limit the diversity of perspectives and introduce interpretive bias. Second, the FW method identifies plausible chains of consequences rather than estimating their likelihood or magnitude. Thus, the results should be interpreted as structured foresight propositions rather than predictive claims. Third, the analysis also focused exclusively on end user perspectives. This limited scope leaves out potential consequences affecting other stakeholders such as developers, regulators, or health care providers. Given the complex nature of implementation pathways, this approach may limit the understanding of the long-term outcomes. Finally, the described effects may not align with the described time horizon of 5 years. AI development speed predictions are highly volatile. Some projections may materialize sooner, while others may never materialize.

Our analysis systematically explores the multifaceted effects of introducing AI into wellness platforms. The projected futures are not uniform and positive; there are parallel trajectories as well as competing ones. Empowerment may arrive together with risk amplification. Better data governance and user-centric design are needed to protect user autonomy. AI wellness apps are aimed at changing individual behavior, but they operate in a broader social and technological context. Complementary future (and futures) studies should include diverse stakeholders and quantify the trade-offs of these platforms.