Gen Z leads in generative AI use overall. But in depth of workplace-specific use, millennials often outpace them. Gen Z experiments more and they’re not afraid of the technology. The root cause is more specific than that.

For Gen Z, AI is often a competitor that may have already taken their first job. That asymmetry of exposure explains most of the behavioural differences we’re seeing.

Dr. Xenia Wade  ·  The Human Side of AI at Work

Gen Z entered the workforce as AI was automating the roles they were supposed to start in. Entry-level hiring at major tech companies has dropped more than 50% over three years, according to SignalFire data reported by Rest of World. LinkedIn, Indeed, and Eures recorded a 35% decline in junior tech positions across major EU countries in 2024 alone.

And the job market is only part of it. On top of job insecurity, Gen Z came of age through multiple serious disruptions: the pandemic, remote schooling, economic instability, and now an AI wave that arrived before they’d had a chance to build a career foundation. That accumulation matters. It shapes how uncertainty feels, and how much of it you can absorb before it becomes resistance.

A 22-year-old watching their first job category disappear isn’t being dramatic. They’re reading the room accurately. No amount of “AI will create new jobs” messaging lands when you can’t get your first one.

This is a question at the heart of generational differences in AI adoption at work, and one I spent years on in my PhD research.

There’s evidence in both directions. Some studies find that the differences we see across generations persist even when you control for age and career stage, which suggests there might be something real about shared formative experiences. Gen Z growing up through Covid, economic instability, and rapid digital change is a genuine cohort experience, not just a life-stage effect. But the more striking finding is that when you account for career stage and role type, the generational signal largely disappears. It’s more of a life-stage difference than a cohort effect. 

My own work on intergenerational knowledge transfer found that what looks like a generational pattern is often better explained by the age difference and also where someone sits in their career. For example, older workers bring deeper contextual judgment due to more experience (Schmidt & Muehlfeld, 2017).

What we’re seeing in the Gen Z sabotage data is most likely several effects at once: career stage, a genuinely disruptive formative context, and a period where AI is right now eliminating the roles they were supposed to start in. The research doesn’t let us cleanly separate these. Anyone who tells you it’s purely a generation thing is oversimplifying.

Dr. Xenia Wade  ·  The Human Side of AI at Work

The practical implication: “what generation are they?” is a less useful question than “what’s their relationship to career risk, and do they trust how this is being introduced?”

My research also points toward something directly relevant here: knowledge doesn’t only flow from older to younger workers; it goes both ways. The types of knowledge exchanged differ, as do the factors impacting the transfer. Applying these findings to an AI rollout, younger employees often bring practical tool fluency, while older employees provide the institutional and social context that determines whether those tools are used effectively. When adoption is structured as a one-way training cascade, both directions of that exchange are lost.

What the Sabotage Is Actually Communicating

The word makes it sound more calculated than it usually is.

Of the workers who admitted to undermining AI strategy, 30% cited fear of job displacement. Another 26% said AI has diminished their value or creativity. These aren’t the same problem (Writer & Workplace Intelligence, 2026).

They are experiencing fear of becoming obsolete, and also likely some type of Identity Drift. It’s the experience of watching the capabilities you built your career on become automatable before you’ve had the chance to build the judgment and seniority that would make you harder to replace.

The workers most afraid of being automated are engaging in the behaviour most likely to get them cut. That’s the Silent Resistance loop: anxiety drives avoidance, avoidance creates a performance gap, the gap increases the very risk they were trying to avoid.

The mistake, when looking at generational differences in AI adoption at work, is to focus on the generation and miss the mechanism.

The mechanism is psychological safety, trust in leadership, and perceived fairness of the rollout. Those dynamics play out differently across generations and career stages, but they affect everyone.

A 2025 peer-reviewed study in Humanities and Social Sciences Communications traced the psychological consequences of AI adoption across 381 employees. The findings: AI adoption erodes psychological safety, which in turn increases employee depression. When leaders show transparency and genuine concern during the transition, the negative effect is significantly reduced.

The WEF’s Future of Jobs Report 2025 found 39% of existing workplace skills will be disrupted within five years. That applies to a 45-year-old just as much as a 25-year-old. Some might quietly avoid the tools, smile in the all-hands, and go home and worry. That’s Silent Resistance, and it’s often harder to see and more costly in the long run.

Each generation is experiencing the same transformation through a different lens. The intervention that works for a 24-year-old facing a disappearing career ladder is not the same one that works for a 52-year-old whose hard-won expertise is being automated. But the underlying need is the same: to feel that this change isn’t being done to them.

Dr. Xenia Wade  ·  The Human Side of AI at Work

1. Separate the fear from the strategy critique

The 30% resisting out of job fear and the 26% who feel AI has diminished their value need different responses. One needs honesty about what’s changing in their role. The other needs to feel that AI is making their contribution more visible, not less. Treating them as one problem fails both groups.

2. Measure where resistance is concentrated

Activation rates tell you nothing useful. The Organisational Adoption Profile measures five readiness drivers (Psychological Safety, Adaptability Mindset, Empowerment Orientation, Action Style, and Adoption Capacity) at the team and role level, so you can see where adoption is actually stalling and why.

3. Build psychological safety before training

Kim, Kim & Lee (2025) show the mechanism: when AI adoption outpaces psychological safety, depression might follow. People need to feel that admitting confusion and making mistakes is safe before they’ll genuinely engage with learning new tools. Training in a low-safety environment doesn’t stick. You need to build psychological safety first.

4. Design for knowledge exchange, not just training cascade

Younger workers often hold tool fluency. Older workers hold institutional context and risk judgment. An adoption strategy that only trains individuals, rather than creating structured conditions for these different kinds of knowledge to meet, leaves significant value on the table.

If people are quietly working against your rollout, the problem starts before training.
The free field guide walks you through three moves to get adoption moving.
→ Get the field guide