Digital mental health has become remarkably good at detecting signals. Stress can be inferred from heart rate variability. Changes in sleep patterns can indicate deteriorating wellbeing. Across the industry, the direction is clear: better detection, earlier intervention, and more personalised support.
The assumption underpinning much of this innovation is that greater visibility leads to greater self-awareness. But that relationship may not be as straightforward as it appears.
Recent advances in AI and wearable technologies have made it possible to quantify aspects of emotional experience that were previously difficult to capture outside clinical settings. Reviews of AI-enabled mental health tools suggest they can improve access to support and help users identify patterns in mood and behaviour (Casu et al., 2024). Similarly, wearable technologies are increasingly being used to monitor stress, recovery, and wellbeing through passive physiological data collection (Kapogianni et al., 2025).
The question for the sector is no longer whether these technologies can support mental health outcomes, but instead what happens when emotional awareness is filtered through them.
Many digital mental health products are designed around a familiar cycle: detect, interpret, intervene. A physiological signal is collected, a state is inferred, and a recommendation is delivered. From a product perspective, the model is attractive. It reduces friction and creates clear moments for intervention.
Yet emotional regulation is not always a problem of insufficient information.
While digital mental health tools are often designed to provide users with more information about their emotional states, evidence suggests that awareness alone does not necessarily translate into effective emotional regulation. Research on interoception and emotion regulation indicates that psychological wellbeing depends not only on detecting internal signals, but also on the ability to interpret and respond to them appropriately (Khalsa et al., 2024; Chen et al., 2025).
This distinction matters because emotional awareness is not purely cognitive. Changes in breathing, muscle tension, energy levels, posture, and movement often precede conscious interpretation. The ability to recognise and make sense of these signals, commonly described as interoceptive awareness, is increasingly recognised as an important component of psychological wellbeing.
Interestingly, some of the most promising research is beginning to move away from the idea that technology should interpret emotional states on behalf of users. Instead, researchers are exploring how digital systems can strengthen awareness of bodily signals themselves. A 2024 study published in Scientific Reports found that technology-mediated sensory feedback could enhance aspects of interoceptive awareness, suggesting a different role for digital mental health tools: not as interpreters of experience, but as facilitators of attention (Goral et al., 2024).
This distinction may become increasingly important as generative AI becomes integrated into mental health products. The value of these systems may not lie solely in their ability to identify emotional states with greater accuracy. It may lie in how they shape the user's relationship with those states.
A system that tells a user they are stressed is solving one problem. A system that helps a user recognise stress before it needs to be labelled may be solving a different, and potentially more important, one.
For digital mental health companies, this raises a design challenge that extends beyond model performance or predictive accuracy. As emotional technologies become more sophisticated, success may depend not only on what systems can detect, but on whether they strengthen the user's capacity for self-awareness in the process.
The future of digital mental health is unlikely to be defined by how effectively technology interprets emotion. It may be defined by how effectively it helps people do that for themselves.
🪺