IT × Well-being: How Can Information Technology Promote Well-being, Not Exacerbate Involution(over competition)?

Understanding and Reflections on Existing Research

Recently, I’ve been immersed in literature exploring how information technology (IT) can promote human well-being, particularly in the sub-field of “health behavior intervention.” The goal of this field is clear: to use technologies like mobile apps and wearable devices to help people cultivate healthy habits (e.g., regular exercise, balanced diet) or quit unhealthy ones (e.g., smoking, excessive drinking, frequent phone checking).

Traditional intervention methods often seem somewhat “crude,” such as sending reminders at fixed times daily, urging you to get up and move. However, this “one-size-fits-all” approach is often ineffective because it ignores the user’s specific context—whether it’s convenient to move, or if they truly need the reminder. Imagine receiving an exercise reminder during a tense meeting or when you’re physically and mentally exhausted; you’d most likely feel annoyed or simply ignore it.

To overcome this limitation, “Just-In-Time Adaptive Intervention” (JITAI) has emerged as a current research hotspot. As the name suggests, JITAI’s core philosophy is to make the intervention system more “intelligent,” capable of providing the “right help” at the “right time.” Here, “right” is not static; it requires the system to perceive and understand the user’s current state and environment in real-time.

So, how does JITAI accurately grasp the “right time”? It primarily relies on two pillars: Context-Awareness and Receptivity.

• Context-Awareness: This means the system needs to know “where you are, what you are doing, and what your surroundings are like.” For example, using phone GPS to determine if you’re at home or the office; using wristband sensors to identify if you’re walking or sitting; or even integrating with weather APIs to know if it’s suitable for outdoor activity.
• Receptivity: Understanding the context isn’t enough; the system also needs to judge “whether you are available to be disturbed and willing to accept suggestions.” For instance, even if it detects you’ve been sedentary for a long time, if you’re busy on a call or feeling down, pushing an intervention message at that moment is likely counterproductive. Therefore, an ideal JITAI system attempts to predict the moments when the user is most likely to accept advice and chooses to intervene then.

To achieve these goals, researchers widely use various types of data. On one hand, there is passive sensing data, which is information automatically collected by devices without user action, such as steps and heart rate from a wristband, or location and activity state (still, walking, driving) from a phone. On the other hand, there is active sensing data, most commonly in the form of “Ecological Momentary Assessment” (EMA), where the system occasionally pops up short surveys asking about your current feelings, stress levels, or cravings (e.g., for a cigarette). Many studies tend to combine these two data types for a more comprehensive picture of the user’s state.

After acquiring data, how does the system decide when to intervene? Currently, rule-based systems are mainstream,
which involves pre-defining a series of “if… then…” logical judgments. For example, “If the user’s step count today is below target, and the current time is afternoon, and the weather is pleasant, then send a walk reminder.” Of course, more advanced research is exploring the use of Machine Learning (ML) or Artificial Intelligence (AI) to aid decision-making. For example, using ML models to predict the user’s likelihood of engaging in physical activity soon, or to assess the user’s current receptivity to intervention messages, thereby more intelligently deciding whether to push a message and what content to send.

In summary, researchers are striving to make health intervention technology “understand you better,” providing personalized support at the most necessary and appropriate times. Although challenges remain, such as real-time stress monitoring and balancing data collection with user privacy and burden, JITAI undoubtedly presents a blueprint for more intelligent and effective health promotion.

When “Optimization” Meets “Involution”: Why Is “Restoration” Over “Competition” a Crucial Goal?

As a master’s student exploring the intersection of Information Technology and Human Well-being (IT × Well-being), I am full of hope for this field’s prospects. By leveraging passive sensing data from mobile and wearable devices (like GPS location, screen time, app interactions) combined with the power of machine learning, we seem poised to understand and even intervene in individual behaviors to an unprecedented degree, aiming to enhance happiness and health. From identifying depression risks and aiding smoking cessation to encouraging exercise and reducing sedentary time, the potential of technology appears boundless.

However, as I delve deeper into the literature and my own reflections, a sense of unease has begun to spread. This feeling intensifies, especially when I encounter studies aimed at optimizing “daily” behaviors—such as using phone sensor data to predict and intervene in academic procrastination, or attempting to reduce “excessive” phone use. These studies often carry an implicit assumption: that successfully changing these so-called “bad” behaviors is equivalent to promoting well-being. But I’ve started to question: Does the “optimization” we so diligently pursue ultimately lead to genuine well-being, or does it, inadvertently, add more fuel to the “involution” (Involution / 内卷) of our times?

Consider this scenario: an app is created that can accurately predict and help students overcome academic procrastination. It might help students reduce anxiety and improve grades in the short term. But what happens if such a tool becomes highly efficient and widespread in an already fiercely competitive academic environment? The likely result is that all students, enhanced by technology, become more “efficient.” But this is followed by greater academic pressure and less free time. The final level of well-being experienced does not increase; it might even decrease. Isn’t this a vivid depiction of “involution”? Technology, in this case, seems to be playing the role of accelerating an “arms race.”

At this point, a critical distinction deserves our deep consideration: why do technological interventions aimed at helping depression patients recover, or assisting people to quit smoking or lose weight, not feel so “involutionary”? I believe the core difference lies in whether the intervention’s goal is “Restoration/Prevention” or “Optimization/Competition”. Behind this, there is a multi-dimensional, deeper logic:

• Psychological Perspective: Restoring Baseline vs. Performance Enhancement. For recognized mental health challenges like depression, the goal of intervention is to alleviate suffering and restore function. It aims to bring the individual from a psychological health level significantly below the norm back up to a functional baseline. This aligns with the logic of treatment, focusing on meeting fundamental psychological needs (e.g., autonomy, competence, relatedness—per Self-Determination Theory, SDT), reducing negative emotions, and restoring their ability to participate actively in life. In contrast, optimizing non-pathological behaviors like “procrastination,” especially in academic settings, often aims to increase efficiency and surpass existing performance, carrying a clear color of enhancement. If this enhancement is driven primarily by external pressures (like GPA, rankings) rather than intrinsic interest or values, then even if efficiency improves, it may not bring genuine psychological satisfaction due to sacrificed autonomy and increased stress.

• Sociological Perspective: Individual Support vs. Systemic Competition. Society generally recognizes the value of providing support to individuals in difficult circumstances (e.g., illness, poverty), seeing it as a duty to promote social equity and overall well-being. Helping a depression patient recover can be understood as helping them regain an equal opportunity to participate in social life. However, technologies like academic procrastination optimizers operate directly within an established, often zero-sum or near-zero-sum competitive system (e.g., a limited number of high-score spots, university admissions). When technology is used to boost an individual’s relative performance within such a system, it is no longer just a personal efficiency tool; it is invisibly changing the competitive landscape. If some gain an advantage through technology, others, to avoid falling behind, may be forced to invest more resources (time, money, energy) to catch up, thus triggering systemic “involution”. The competition becomes fiercer, but the overall “pie” doesn’t get bigger, the individual’s return on investment drops, and so does their well-being.

• Philosophical & Ethical Perspective: The Justification of Treatment vs. The Controversy of Enhancement. Ethically, “treatment” aims to fix defects and restore health, and is generally considered to have strong moral justification, often rooted in respect for life and compassion for suffering. “Enhancement,” however, aims to transcend so-called “normal” standards and improve abilities, and its ethical boundaries are much blurrier. The use of enhancement technologies in competitive fields, in particular, easily sparks profound controversies about fairness, identity, and even “what it means to be human.” Helping a student eliminate procrastination to get higher grades focuses more on instrumental value (serving some external goal), unlike treatment, which has a stronger intrinsic value (health itself is the goal). An excessive focus on instrumental optimization can also lead to the alienation of humans—where individuals are seen as “productivity units” needing constant tuning, rather than as complete beings with rich inner experiences and diverse values.

• In short, “restoration/prevention” interventions aim to help individuals reach a socially recognized baseline of health or function. They are more of a form of support and empowerment and are less likely to directly intensify existing competition. Conversely, “optimization/competition” interventions, when applied in an already high-pressure competitive environment, are highly likely to exacerbate systemic involution by upgrading individuals’ “equipment,” ultimately running counter to the original intention of improving well-being.

Beyond Behavior Modification: Towards an Integrated, Intelligent, and Humanistic Well-being Support System

Recognizing the problem is just the first step. More crucial is to ask: can we design a different technological path, one that allows IT to truly serve holistic human well-being, rather than just optimizing single metrics that may worsen involution?

My current thinking is that we might need to shift from a model of “passive monitoring, post-hoc intervention” to a paradigm of “proactive planning, pre-ante support.” In much current research, technology acts more like a “behavioral supervisor,” intervening only when the user commits a “violation” (like excessive phone use). But what we might truly need is a role more akin to an intelligent “life navigator” or “executive function assistant.” This requires technology to have a more fundamental, deeper understanding of the user, beyond just keeping a log of behaviors.

Imagine the data structure needed to support such a novel system. It would need to include these key informational dimensions:

1. User Goals: No more vague assumptions that the user wants to “use their phone less.” It requires explicitly recording and understanding the user’s diverse goals:

   • GoalID: G001, Description: "Complete thesis", Type: Academic, Priority: High, Deadline: "2025-12-15"
   • GoalID: G002, Description: "Work out 3 times a week", Type: Health, Priority: Medium
   • GoalID: G003, Description: "Keep in touch with friends", Type: Social, Priority: Low
   • It even needs to understand goal hierarchies, e.g., G001 contains sub-goal "Complete literature review".

2. Activity Library: The system must understand the nature of different activities and their potential impacts:

   • ActivityID: A01, Description: "Deep work (thesis writing)", EstimatedDuration: 120min, AssociatedGoal: G001, RequiredState: {Focus: High, Energy: Medium}, PredictedImpact: {Goal_G001_Progress: +15%, Energy_State: -2}
   • ActivityID: A02, Description: "Go to the gym", EstimatedDuration: 90min, AssociatedGoal: G002, PredictedImpact: {Energy_State: -3 (short-term), +2 (long-term), Mood: +1}
   • ActivityID: A03, Description: "Browse social media", EstimatedDuration: 30min, AssociatedGoal: None (or weak link to G003), PredictedImpact: {Focus_State: -2, Mood: +1 (short-term), -1 (if it cuts into G001 time)}, IsPotentialDistraction: True

3. User State: Real-time or near-real-time sensing of the user’s internal state, going beyond simple screen-on/off logs:

   • Timestamp: 2025-05-01 19:00, Mood_Pleasantness: 5/10, Energy_Level: 3/10, Cognitive_Load: High, Anxiety_Level: 7/10, Source: EMA survey + Sensor inference

4. Context: External factors like location, time, and environment:

   • Timestamp: 2025-05-01 19:00, Location: Home (at desk), Time_Period: Evening, Day_of_Week: Thursday, Ambient_Noise: Low

5. Intelligent Planning & Intervention Logic: By integrating all the above information, the system can make truly intelligent judgments and recommendations. For example:

   • Scenario: User (State: low energy, high anxiety) opens a social media app (Behavior Log) in the evening (Context: home, evening).
   • Traditional Intervention: “You have been using your phone for over XX minutes!” (Likely ineffective, may cause resistance)
   • New System Intervention: “Noticed you might be feeling a bit tired and anxious. Browsing social media right now might make it harder to focus on your thesis later (Goal: G001). Would you like to try relaxing for 15 minutes with some calming music (Activity: A04 - Relaxation) first? Or maybe do some simple thesis-related file organization (Activity: A05 - Low-cognitive-load task)? This might help you recover some energy and make a little progress on G001.”

Why is such a data structure necessary? Because only then can the system:

• Understand the “why” behind the behavior: Is the user actively seeking relaxation, or are they escaping a difficult task?
• Assess the “opportunity cost” and “long-term benefit” of the behavior: What is the real impact of the current action on long-term goals and short-term state?
• Provide personalized, contextual, and goal-oriented advice: No more “one-size-fits-all” blunt reminders, but rather constructive, supportive feedback based on the user’s current state and long-term aspirations.

The Long Road Ahead: Challenges and Hopes

Achieving such an idealized system undoubtedly faces enormous challenges. It requires significant breakthroughs in the precision and non-intrusiveness of data collection, the accurate inference of user states, the effective modeling of personal goals, the dynamic prediction of behavioral impacts, the efficiency of intelligent planning algorithms, and the natural fluency of human-computer interaction. More importantly, we must be extremely cautious in handling the accompanying privacy and ethical considerations.

But this is perhaps the necessary path for IT × Well-being technology to return to its original purpose of promoting human flourishing. We cannot be satisfied with merely optimizing singular behavioral metrics; we must commit to building technological systems that can understand and support the complex inner needs and long-term values of individuals.

Although not as directly impactful on social productivity as fields like energy science or AI, IT × Well-being might help individuals find more breathing room in a society full of competition and suppressed needs. And when more researchers and users think about this problem—“promoting well-being, not fueling involution”—perhaps that in itself is a contribution to solving it.

Returning to Myself: My Struggles, Reflections, and Next Steps

So, bringing the focus back to myself—a novice learner in this field—what can I do?

Frankly, a large part of what drives me to ponder these grand questions stems from my personal experience: a long-term coexistence with Obsessive-Compulsive Disorder (OCD) and Generalized Anxiety Disorder (GAD). Initially, I had a simple impulse: to develop a tool that could help me, and perhaps others like me who struggle in the quicksand of self-regulation, to find a moment’s relief from endless anxiety and procrastination.

But as discussed earlier, this idea quickly threw me into a deep ethical dilemma. Many of my struggles are closely related to past high-pressure academic environments. So, if such a tool, designed to improve “efficiency” and reduce “internal friction,” were placed in a “cram school” like a high-intensity high school, or a university campus with a growing atmosphere of “sophisticated self-interest,” would it not truly degenerate into another “involution-aiding” artifact?

If so, what would be the point of me developing this tool? Is it merely to help myself, or to help other vulnerable (or even more vulnerable) individuals, to alleviate some suffering so that they can… more “effectively” participate in this seemingly endless game of involution? After all, those who are truly in their element, the “involution kings” (卷王), or those naturally blessed with strong self-regulation, might not need such a tool at all. And for those of us who “need assistance,” if we are merely given a stronger “competitive weapon,” does it not, in the end, just intensify the brutal infighting among those in the middle and lower parts of the pyramid?

This thought sends a chill down my spine.

Facing such ethical dilemmas and practical challenges, the “Plan B” I can think of, what is within my power, is perhaps this: to stop focusing solely on developing a perfect “solution” product, and instead, treat the entire process of exploration, reflection, connection, and sharing as a more meaningful form of action. Instead of fixating on a potentially “involution-aiding” efficiency tool, I should try to explore a path that truly leads to inner peace and a meaningful life.

Specifically, my next steps might include:

1. Starting from “Me,” but Connecting to a Broader “We”:

   • Use solving my own procrastination and anxiety as a genuine starting point for self-experimentation. Try to design and conceptualize tool prototypes or core philosophies that support internal needs (not just optimize external performance). Treat this process as an opportunity for deep learning and understanding of the problem.
   • Embrace an open and connected attitude: I will consider open-sourcing my thoughts, attempts, and even my immature code snippets, and actively participate in broader discussions within the technology and mental health communities. I will honestly admit my limitations (e.g., my development skills are very limited!) and actively seek collaboration and interdisciplinary help.
   • Small-scale co-creation and in-depth interviews: Invite friends I know who also feel the need for similar “assistance” to join as early participants. The goal is not simple “user testing,” but to jointly explore and define “What is truly helpful support?” and to understand the specific needs and concerns of different individuals in their struggles.

2. Exploring Roots Inwardly, Disseminating Insights Outwardly:

   • Continuous learning and deep thinking: Read more broadly in related fields like psychology, sociology, and philosophy, especially on deep topics like self-regulation, motivation theories, the science of happiness, meaning in life, and building personal values. Continuously explore inwardly, striving for a deeper understanding of the problem’s roots, rather than just its behavioral surface.
   • “Translation” and popularization of knowledge: Try to take the professional theories, excellent research, or profound articles I learn and popularize, interpret, and share them in an more accessible way. I believe that raising the public’s cognitive level and depth of discussion on issues like mental health, well-being, and tech ethics is an extremely valuable action in itself.
   • Advocating reflection and initiating discussion: In my content sharing, actively integrate reflections on the “involution” phenomenon and the socio-cultural factors behind it. Advocate for more diverse, humanistic evaluation systems and lifestyle choices. At the same time, actively initiate or participate in related discussions, to think together with more people: In this age of rapid technological change, how can individuals find and establish a sense of meaning, control, and value in life?

I am well aware that this path is full of unknowns and challenges, and there will be no easy answers along the way. But perhaps, it is in this very process of constant reflection, learning, connecting, and trying amidst confusion that the potential for a more authentic and sustainable well-being lies.

Ultimately, I hope my humble efforts, however small, can contribute to that grander goal: to make technology truly serve the richness of humanity, helping us live more freely, more authentically, and more meaningfully—not just more “efficiently” or more “competitively.” Because I believe that here, “the free development of each is the condition for the free development of all.”