Opinion: Mental health in the age of Algorithm

It isn’t AI alone that hurts employees; it’s how it is deployed, governed, and understood as part of organisational environment

By Akshita Pandey, Dr Moitrayee Das

While companies highlight AI’s advantages: greater efficiency, fewer errors and data-driven precision, there is a vital and usually overlooked expense: the effect of AI implementation on workers’ mental well-being.

---

An increasingly large body of scholarship is starting to uncover the layers of life and work under algorithmic rule. Among the strongest additions to this discussion is a 2024 paper by Byung-Jik Kim and Julak Lee, in Humanities and Social Sciences Communications. Based on a survey of 418 South Korean professionals, one of the country’s most technologically advanced populations, the research examines the impact of AI adoption in the workplace on worker stress, burnout, and psychological well-being.

Invisible Weight of Algorithms

The study reveals a psychological domino effect: AI adoption does not necessarily lead to burnout, but it drastically increases job stress, which becomes a significant predictor of burnout. It’s a subtle move with far-reaching implications. Workers are not necessarily angry about the existence of AI itself. Instead, it’s the pressure AI imposes upon them—the rate of change, the ongoing requirement to learn, the threat of obsolescence—that stokes this growing worry.

AI fosters an environment of techno-stress—a catch-all term for several stressors that include:

● Techno-overload: being overwhelmed by the compulsion to deal with new digital tasks and devices

● Techno-insecurity: the fear that automation will take one’s job

● Techno-complexity: having difficulty learning and adapting to AI systems promptly

● Techno-uncertainty: uncertain expectations regarding performance under AI systems.

This pressure comes at a high psychological cost. Tasks that once relied on human judgment and creativity are now often shaped by algorithms, leaving employees feeling less autonomous and more like extensions of the systems they work within. This shift has contributed to rising feelings of emotional exhaustion, disengagement, and helplessness, a pattern reflected across industries ranging from finance and education to healthcare.

Psychological Theories

Kim and Lee base their conclusions on three dominant psychological paradigms:

● The Job Demands-Resources (JD-R) Model: Proposes that burnout occurs when job demands (such as AI upskilling) exceed available resources (eg, training, assistance, or affective resilience).

● The Conservation of Resources (COR) Theory: People attempt to preserve their physical, emotional, and psychological resources. As AI compromises these resources, eg, by making employees feel irrelevant or incompetent, people suffer from stress and burnout.

● The Transactional Model of Stress and Coping (TMSC): If AI is seen as a threat instead of an opportunity, and if coping skills are poor or lack support, then the outcome is psychological strain.

Simply put, it isn’t AI alone that hurts employees; it’s how AI is deployed, governed, and understood as part of an organisational environment.

Digital Self-Efficacy

Another influential finding is AI learning self-efficacy, or in simple terms, an individual’s confidence in their ability to master and utilise AI. Such employees feel less stressed and more resilient when encountering technological change. Low self-efficacy results in helplessness, anxiety, and burnout.

In this context, self-efficacy represents a new form of workplace armour. It does not merely enable employees to excel; it equips them to cope emotionally in AI-powered work environments.

New Face of Burnout

Historically, burnout has been linked to long working hours, poor management, unrealistic expectations, and emotionally demanding jobs. However, in the algorithmic era, burnout is taking on a new and more invisible form, one rooted not only in overwork, but also in disconnection, disempowerment, and constant digital overload.

Today, algorithms shape how we work, communicate, consume information, and even rest. Notifications blur the boundaries between personal and professional life, while productivity-driven digital cultures create pressure to always be available, responsive, and efficient. At the same time, algorithmic systems reduce autonomy by monitoring performance, predicting behaviour, and dictating routines in ways that can make individuals feel replaceable and emotionally detached from their work. The result is a form of burnout that is less about physical exhaustion alone and more about cognitive fatigue, emotional numbness, and the persistent feeling of never truly being “offline.”

In the algorithmic era, burnout is taking on a new and more invisible form, one rooted not only in overwork, but also in disconnection, disempowerment, and constant digital overload

Employees increasingly report feeling sidelined by AI systems that make decisions with little to no human input. Others describe a growing sense of being constantly “watched” through performance metrics, productivity trackers, and algorithmic evaluations that monitor everything from response times to efficiency levels.

This culture of continuous surveillance often leads employees to engage in chronic self-monitoring, creating heightened pressure to appear productive at all times. Over time, this can contribute to anxiety, emotional exhaustion, and a weakening sense of personal agency within the workplace. This type of burnout is less obtrusive, less visible, but no less harmful. It may undermine self-esteem, drive up absenteeism, and, in extreme instances, result in depression or leaving the workforce altogether.

What Can Be Done

Normalise psychological support in AI transitions: Organisations need to realise that implementing AI is not merely a technical change but a psychological one. That requires providing mental health support, check-ins, and safe spaces for employees to vent anxiety or confusion.

Foster digital self-efficacy across the board: Continuous training, mentoring, and peer support initiatives to develop digital confidence are a must. The objective is not technical competence alone but emotional resilience.

Redesign AI with empathy in mind: AI systems must be constructed with user agency at their centre. Workers need to know how algorithms work, what data is being drawn upon, and where human input is placed in the decision-making pipeline. Transparency and inclusion lower fear and resistance.

Reframe measures of success: Human feedback, teamwork, and imagination need to be preserved in performance measurement.

Create ethical principles around mental health and AI: Mental health needs to be an integral part of ethical AI, no less than privacy, bias, or transparency.

Re-Centering the Human

One of the most significant contributions of Kim and Lee’s research is that the real cost of AI isn’t simply in terms of job loss or data risk, but in the emotional lives of actual human beings. AI isn’t neutral. AI alters how people feel about their work, their worth, and their future. If we forget the mental health consequences of AI adoption, we risk creating an efficient but inhuman workplace, optimised but emotionally unsustainable.

The algorithm age is upon us. The question is whether we will greet it with empathy, foresight, and care, or whether we will trade human well-being at the altar of automation. The future of work doesn’t just rely on smarter machines. It relies on more humane systems.

(Akshita Pandey is undergraduate student, and Dr Moitrayee Das is Assistant Professor of Psychology at FLAME University)