Is AI Making Us Lonelier at Work?

Source: Pexels / Energetic

As more companies take an AI-first approach, increasing interactions with AI systems could potentially heighten loneliness and social isolation, particularly for those who already experience social anxiety. A set of studies published in the Journal of Applied Psychology reveals that AI systems in the workplace have unintended emotional consequences of increasing loneliness and social disconnection.

These studies were conducted in four countries (Taiwan, Indonesia, Malaysia, and the United States) and analyzed interviews and third-party observations from coworkers and family members of over 794 workers who frequently interact with AI systems for work. Each study examined different professions, ranging from biomedical engineers to real estate consultants to those in marketing, finance, and accounting.

Researchers found that the more employees collaborated with AI, the more they felt socially deprived and lonely, particularly for those who already experienced attachment anxiety. Working with AI throughout the day led to:

• A greater need for social connection
• Increased loneliness and social isolation
• Higher rates of insomnia and alcohol use after work
• Stronger effects in individuals with attachment anxiety

Even though the loneliness drove people to connect with human coworkers more (prosocial compensation), participants still reported feeling socially isolated even after these attempts to connect. These findings have important implications for both mental health clinicians and organizations.

We Are Wired for Connection, Even At Work

Humans are biologically wired for connection. Our nervous systems evolved in small groups where cooperation and social cues (facial expressions, eye contact, body postures, and gestures) were essential for survival.

When those cues are absent, as they often are in AI interactions, our innate regulatory systems flag a social deficit. This triggers our “need for affiliation,” a drive to connect and have a sense of belonging. If we do not get this need fulfilled, it creates a chronic "social deficit," which can lead to anxiety, lower mood, and disengagement. The resulting loneliness around increasing AI interactions at work can be deep and unconscious, showing up as insomnia, demotivation, resignation, numbness, detachment, disconnectedness, or disillusionment.

Researchers found that some workers responded adaptively by helping others, which may have helped restore a sense of belonging. But others experienced a darker side of this deprivation instead: increased loneliness. This loneliness did not end when the workday was over. It impacted personal life, with increased self-reported rates of alcohol use and observed rates of insomnia.

Workers With Attachment Anxiety May Experience Greater Loneliness

The study suggests that AI interactions disproportionately affect workers with higher levels of attachment anxiety, a style of relating to others characterized by a strong desire for closeness along with a fear of abandonment and rejection. For these individuals, daily interactions with AI can amplify feelings of isolation, abandonment, and social inadequacy.

Tang PM, et al. No person is an island: Unpacking the work and after-work consequences of interacting with artificial intelligence. J Appl Psychol. 2023.

Source: Tang, et. al (2023)

For clinicians and organizational leaders alike, it is important to consider AI systems as not just tools, but also "coworkers" or social agents with real psychological impacts on the rest of the team.

Takeaways for Mental Health Clinicians and Leaders

With increasing AI interaction at work, there may be unintended emotional side effects on morale and sense of belonging. Many individuals will be able to adapt, finding ways to fulfill their need for social connection in other ways. But others, especially those who experience attachment anxiety, may find themselves lonelier.

An AI-integrated workplace will necessitate more psychological resilience and flexibility. Here are some clinical and organizational takeaways:

1. Recognize AI-mediated loneliness to help reduce feelings of alienation and isolation. Self-compassion and recognition of the feeling can be an antidote to the sense of isolation and invisible loneliness. Paradoxically, the recognition that we are not alone in this experience of alienation can heal some of it.
2. Explore more experiences of meaningful human connection to offset workplace loneliness. Organizations may be able to build, proactively, more moments of peer collaboration into the workflow. For those feeling less of a sense of belonging at work, finding other places for community can be helpful.
3. Signs of increased anxiety and loneliness may signal a need for more support and may have deeper roots. Individuals with chronic loneliness and interpersonal anxiety are most at risk. Certain types of psychotherapy, including relational, psychodynamic, and experiential work, can go deeper to help people redo early formative experiences and heal original attachment wounds.
4. Psychotherapy, rooted in human relationships and delivered by human therapists, will remain essential for healing deeper loneliness. Psychotherapies rooted in the relationship and that offer an authentic experience of another human being are key to healing deep loneliness.

Artificial Intelligence Essential Reads

How Do Human Beings Stay Relevant in an AI-Fueled Future?

AI, GLP-1s, and the Fear of Lazy Shortcuts

The likely correction for AI-mediated workplace loneliness will be deeper human connection rather than designing AI systems at work to be more socially engaging (e.g., AI work companions). The dangers of a more socially engaging AI at work could lead to a different set of problems, such as emotional dependence, overreliance, overworking, or bias in strategic decision-making.

AI systems or "AI coworkers" can be enormously helpful, including research showing that AI "teammates" help increase productivity and broaden core expertise. But such technological implementation is not always emotionally neutral. This does not mean limiting AI implementation, but it is helpful to consider the psychological consequences since the short- and long-term social and emotional impact is currently unknown.

Marlynn Wei, MD, PLLC © Copyright 2025 All Rights Reserved.