If employers across the economy are to prevent a damaging, AI-driven “anticipatory exodus” of white-collar workers, they will need to bridge the “confidence gap” and actively mitigate fears that such jobs are surplus to requirements, experts say.

As Tina Shah Paikeday, a Researcher at the Drucker School of Management at Claremont Graduate University in California, points out:

The anticipatory exodus is driven by fear, rather than a displacement exodus driven by actual job loss. White-collar workers are trying to understand whether AI will make their roles more valuable, less secure, or simply harder to navigate. That uncertainty can absolutely affect retention, career planning, and willingness to stay in certain fields.

A key problem here, she says, is that:

AI capabilities have moved faster than the change management, workforce planning, management, and training systems most organizations had in place. As a result, employees often feel the uncertainty before employers have clear answers. That creates risks for engagement, retention, skills development, and productivity. For the wider economy, the challenge is a potential mismatch between where AI can drive gains and where workers feel confident enough to stay, learn, and adapt.

While it is not too late to address the situation, Paikeday believes:

The window for getting ahead of it, rather than cleaning up after it, is beginning to close. Employers need to move from vague reassurance to practical workforce design. That means explaining which tasks AI will support, which decisions still require human accountability, what new skills employees need, and how career progression will work as roles evolve. 

AI as a workforce transformation initiative

In other words, she says, employers need to treat AI as a workforce transformation initiative, rather than simply a technology rollout, to ensure they are “much better prepared”. Unfortunately though, Paikeday adds:

I’d characterize employer readiness broadly as behind the anxiety curve — workers are feeling the uncertainty before most companies have articulated a clear plan.

Joel Marotti, Senior Managing Partner at career services provider Vertical Media Solutions, certainly agrees that the time to treat AI purely as a technology initiative is over. In his view though, it should be looked at more like a “retention conversation”:

Every knowledge worker is privately calculating whether their job is even going to exist in three years. If the company isn’t getting in front of this and openly addressing it, these folks are just answering it alone, and it’s not usually going to be the most positive answer there. 

A second consideration, Marotti says, is the importance of investing in “transition skills” rather than just tools. For instance, he points out:

Only 35% of workers over 50 say their employer is doing enough to train them on AI, and that gap between anxiety and support is really where you’re losing a lot of people. So, the third point is to create more clarity. One of the biggest drivers of AI anxiety right now isn’t that people think robots are coming. It’s really that nobody has told them what their job is going to look like in 18 months. And I think that silence is causing people to start thinking, and that’s driving a mass exodus.

AI as a path to more valuable work

On the other hand, warns Paikeday, simply treating employee concerns as if they are a “communications issue” once the real decisions have already been made is a recipe for disaster too. Instead, she says:

The organizations that navigate this best will involve employees early, invest in reskilling, create clear guardrails, and make AI feel like a path to more valuable work rather than quiet replacement…It’s all about disciplined, early knowledge-transfer and reskilling before the expertise walks out the door.

Mohaimen Bayoumi, an HR and Talent Strategy Leader and Future of Work Thought Leader, agrees. He explains how he has dealt with the risk of losing skills and experience in the past:

In one of my previous organizations, we addressed this by strengthening mentoring programs, encouraging cross-generational collaboration, and building structured knowledge-sharing practices so critical expertise didn’t leave with individuals. Going forward, organizations should make knowledge transfer a strategic priority through mentoring, succession planning, and AI-powered knowledge management that captures expertise before it’s lost. Leadership development also needs to start much earlier rather than waiting until employees reach management positions. Organizations that build leadership pipelines early, instead of reacting when vacancies arise, will be far better prepared for the future.

A key consideration here, Bayoumi says, is not treating AI solely as a cost reduction tool:

One of the biggest mistakes organizations can make is viewing AI purely as a cost-cutting initiative. I’ve seen businesses take that approach, only to reverse course after losing capability, productivity, and significant amounts of money. AI should be treated as an augmentation strategy – not a replacement strategy.

AI as a workforce design challenge

As to what the future is likely to hold, Bo Young Lee, Chief Executive of AI4ALL, a non-profit that provides AI learning support to young people aged between 16 and 24, believes:

In the short term, we will likely see continued anxiety, experimentation, and uneven adoption. My concern is that if too few people enter knowledge-based professions because they perceive AI as a threat rather than an opportunity, we could create a significant talent pipeline problem. Organizations may find themselves facing a severe shortage of mid-career and middle-management talent in a decade or so. These are the people who translate strategy into execution, mentor the next generation, and drive much of an organization’s productivity, innovation, and progress.

Over the medium term though her view is that organizations will start redesigning jobs based around human-AI collaboration. As a result, some roles will disappear, others will evolve, and entirely new categories of work will likely emerge too. 

As for longer term outcomes, whether they are positive or negative will depend on the choices employers make today. Lee explains:

If we use AI primarily to reduce labor costs, we risk increasing inequality and weakening pathways to economic opportunity. If we use it to expand human capability and create broader access to expertise, AI could become one of the most powerful tools for economic growth and societal advancement. Responsible AI isn’t just about building better technology but ensuring we don’t automate away the pathways people use to build expertise, economic mobility, and meaningful careers. The future is not predetermined. The question is not whether AI changes work, because it absolutely will. The question is whether we build institutions, educational systems, and organizations capable of helping people adapt alongside it.

Paikeday agrees, adding that in a workplace context:

The strongest employers will be those that treat AI implementation as a workforce design challenge, not simply a software development. AI may create new jobs and productivity gains, but only if works trust the transition enough to participate in it.

My take

It seems we are at a pivotal moment in terms of the impact AI will have on the workplace, and wider society by default, going forward. The future is in the hands of today’s employers and the decisions they make today will reverberate throughout the generations to come, for better or worse. So, to avoid dystopian outcomes, leaders must be careful to put people – rather than technology, efficiency or productivity considerations – at the heart of the matter.