Column
AI Overuse at Work Is a Competitive Environment Problem, Not a Personal Discipline Problem
A study of 396 generative AI users found that social comparison orientation — not individual personality — drives problematic AI use through FoMO and perceived replaceability. The design implication: competitive workplace structures create the trap, and organizational design can dismantle it.
Hi, I’m Keito Inoshita from Affectosphere Group.
A pattern that’s become common in enterprise AI rollouts: make AI usage visible.
Leaderboards showing who’s using the tools. Best-practice showcases spotlighting high-usage employees. Utilization dashboards tracking hours per person. The intent is good — creating social proof and motivation to adopt.
But a paper published on arXiv in June 2026 (Xuchao Zhang, Jihye Lee, arXiv:2606.03560) raises a problem with this approach. Using structural equation modeling on 396 generative AI users, the researchers found that social comparison orientation — the tendency to evaluate oneself relative to others — directly and indirectly drives problematic generative AI use.
Problematic use includes overuse, dependency, and what the researchers describe as workaholism-like AI reliance: being unable to function without AI assistance and feeling anxious about falling behind.
The core finding: this isn’t primarily about individual personality. It’s about competitive social contexts — which organizations are actively designing into their AI rollouts.
3 Points for Today
- The cause is structural, not personal: social comparison orientation, not individual weakness, is the primary driver of problematic AI use. Competitive visibility creates the conditions.
- FoMO is the transmission mechanism: fear of missing out mediates the path between comparison and problematic use — “everyone else is mastering AI, I’m falling behind” turns into overuse and dependency.
- Organizational design is the intervention: competitive environment → collaborative environment is a design change, not an awareness campaign. It can be implemented.
① How Social Comparison and FoMO Drive Problematic Use
Social comparison orientation (SCO) is the tendency to evaluate one’s own abilities and situation by comparing them to others. High-SCO individuals are more likely to feel inadequate when they see others performing well — even if their own performance is objectively fine.
The study found a direct relationship: higher SCO predicts higher problematic AI use scores. But the indirect paths are where the organizational design implications become clear.
Path one runs through FoMO (Fear of Missing Out). When someone with high SCO sees colleagues using AI tools fluently, the response is: “I’m being left behind. I need to use this more.” This FoMO-driven urgency translates into overuse that isn’t motivated by task efficiency — it’s driven by anxiety about relative positioning.
Path two runs through perceived replaceability. “If I don’t demonstrate AI proficiency, I’ll become redundant” is the fear here. The response is defensive overuse: using AI tools extensively not because it’s helping but because not using them feels risky.
Both mechanisms are amplified by environments where AI usage is visible, ranked, or incorporated into performance signals.
② How Workplace AI Rollout Design Creates the Problem
The most important implication of this research is that the organizational design choices made during AI deployment can either suppress or activate these mechanisms.
What increases risk:
Individual usage visibility. Dashboards showing who is using AI tools how often, recognition programs highlighting top users, team meetings where AI adoption is tracked per person — these create comparison contexts directly.
Incorporating AI usage into performance metrics. “Percentage of proposals AI-assisted,” “task throughput using AI tools,” as explicit KPIs make adoption competitive rather than instrumental.
What reduces risk:
Team-level and aggregate visibility instead of individual-level. “Our team’s AI-assisted workflow completion increased 12% this quarter” doesn’t create individual comparison pressure the way “the top five users this month” does.
Learning-oriented sharing formats. “What interesting things did you try with AI this week?” is a different social context than “who is using it most.”
These aren’t minor framing differences. The research suggests they change the social comparison context, which changes FoMO, which changes problematic use rates.
③ What HR and Organizational Development Teams Can Do
Translating the research into practical organizational interventions.
Design AI adoption culture around learning, not competition
If you’re running an internal AI community or best-practice sharing program, structure it around experimentation and curiosity rather than usage quantity. The question “what did you find useful?” creates a different social dynamic than “how much are you using it?”
This is a joint design project between CHRO, organizational development, and the AI deployment team. A measurable outcome: add a “psychological safety around AI adoption” item to your engagement survey and track it quarterly. If competitive anxiety is present, it will show up in the data before it shows up in attrition or performance problems.
Remove individual-level AI usage comparisons from dashboards and meetings
If your current setup includes per-person AI usage metrics visible to the team or manager, consider switching to aggregate tracking. The goal is keeping the organizational-level signal (are we getting more efficient?) without creating the social comparison conditions that drive FoMO.
This is an easy design change that most HR tech and IT teams can implement in a configuration, not a build.
Identify high-FoMO employees as a proactive support opportunity
High-FoMO employees are the highest-risk group during AI rollout periods. In 1:1 check-ins or pulse surveys, a single item — “Do you feel anxious about falling behind your colleagues in AI use?” — can surface employees who need support before that anxiety translates into dependency or burnout patterns.
Positioning this within an EAP (Employee Assistance Program) framework, as a “technology-related workplace stress” category, is one scalable way to operationalize early support.
From “Get People to Use It” to “Help People Use It Well”
AI deployment projects often end up with utilization rate as the primary success metric. Usage dashboards, adoption leaderboards, and AI performance KPIs are the natural outputs of that framing.
But this research suggests that optimizing for utilization rate, through competitive visibility, can generate the psychological conditions for problematic use: overuse, dependency, anxiety-driven adoption that looks like engagement but isn’t.
The alternative isn’t to slow down AI adoption. It’s to decouple AI use from competitive self-evaluation by designing a social context where it doesn’t need to be a signal of status or replaceability.
“How much are we using it” is the wrong question. “Is using it working for us” is closer to what matters — and it requires a different kind of visibility to answer.
That’s all for today!
References
- Xuchao Zhang, Jihye Lee (2026). The Comparative Trap: How Social Comparison Orientation Drives Problematic Generative AI Use. arXiv preprint.
* This article was written in part with AI assistance and may contain inaccuracies.