By Guest commentary Friday, May 22nd, 2026

By Vlad Vaiman

The conversation about artificial intelligence and employment has hardened into two opposing camps. 

Optimists promise a productivity renaissance — a future where AI handles the drudgework and humans are freed to do their best thinking. 

Pessimists warn of mass displacement, of entire professions rendered obsolete, of an economy that simply stops needing as many people. Both camps are telling a version of the truth. 

But neither is telling the whole story.

What’s actually happening inside companies right now is quieter, stranger, and ultimately more consequential than either narrative admits. 

Work isn’t being eliminated so much as it’s being disassembled and reconstructed — tasks reshuffled, roles redefined, value chains rewired. The right word isn’t replacement. It’s reconfiguration.

Start with the statistic that gets cited most often: that more than half of current work activities in the United States could be automated using technology that already exists. 

That number sounds alarming. It’s also misleading.

The problem is that it conflates tasks with jobs, and those are not the same thing. Jobs aren’t neat bundles of interchangeable parts that can be cleanly extracted and handed to a machine. 

They’re integrated, contextual, and deeply social. 

They involve judgment calls made on incomplete information, relationships built over years, and the kind of institutional knowledge that doesn’t fit in a prompt. Pull one thread and the whole fabric shifts.

Software development is the clearest illustration. 

AI can now generate a substantial share of functional code. And yet software developers aren’t disappearing; they’re redirecting. Less time writing, more time reviewing, questioning outputs, integrating components, and aligning what the machine produced with what the system actually needs. 

The work evolves. The worker remains. This isn’t a semantic distinction. It’s the entire argument.

The deeper shift, the one that will define the next decade of business strategy, is in where human value actually lives.

As AI absorbs structured, repeatable, and increasingly complex analytical tasks, the work left for people isn’t the leftovers. 

It’s often the most consequential work of all: judgment under uncertainty, reading a room, navigating ambiguity, making ethical calls, sustaining the kind of trust that takes years to build and seconds to destroy. 

These capabilities aren’t being devalued by AI. If anything, they’re being bid up.

That creates real pressure on workers. 

Baseline competence is no longer a defensible position. 

The people who will thrive aren’t those who resist AI, nor those who simply adopt it, but those who learn to use it as a force multiplier while simultaneously deepening the distinctly human skills that machines can’t replicate: critical thinking, negotiation, collaboration, the ability to ask the right question rather than just process the answer. 

Call it super-skilling. It’s a higher bar than most organizations are currently preparing their people to clear.

For business leaders, the challenge is equally steep, and the failure modes are already visible.

Companies struggling to see meaningful returns on AI investment typically aren’t suffering from bad technology. They’re suffering from piecemeal implementation. 

AI gets bolted onto existing workflows, a tool added here, a pilot launched there, with the underlying structure of work left largely intact. The efficiency gains are real but modest. The transformative potential goes unrealized.

The real returns come when firms step back and ask harder questions – not which tasks can be automated, but how work should be fundamentally reorganized around what AI makes possible. 

That means redesigning processes from the ground up, rethinking roles across entire value chains, and resisting the temptation to declare victory after a promising proof of concept.

It also means being honest about the limits that public discourse tends to gloss over.

One is what researchers from the Center for Effective Organizations at USC Marshall call indivisibility. Most jobs, especially in smaller businesses, are too integrated to break apart cleanly. Automating a slice of someone’s role doesn’t produce a proportional reduction in headcount. 

It produces a modified role with new coordination demands: reviewing AI outputs, catching errors, and translating machine logic into human context. Efficiency in one place creates complexity somewhere else.

Another limit is what might be called transitivity – the assumption that swapping human input for AI output leaves value intact. It often doesn’t. In customer service, healthcare, and professional services, the relationship is frequently the product. 

Clients aren’t just buying an outcome; they’re buying confidence, continuity, and the sense that someone is genuinely accountable. Automate that away in the name of cost savings and you may find you’ve also automated away the reason customers stayed.

None of this is an argument against AI adoption. Businesses that move slowly will fall behind. The technology is too powerful, and the competitive pressure too real, to treat caution as a strategy.

But adoption without intelligence is its own kind of risk. The organizations that will come out ahead over the next decade are not the ones that replace humans most aggressively. 

They’re the ones that integrate humans most thoughtfully, designing systems where technology amplifies what people do best, rather than attempting to route around human involvement altogether.

That requires a different mindset than the one driving most AI conversations right now. 

The technology-first framing – what can AI do, what can it automate, how fast can we scale – needs to give way to a system-level framing: how do we design work, organizations, and human-AI collaboration in ways that actually create durable value?

Three capabilities will define the companies that get this right. 

First, the ability to redesign work at scale, not task by task, but end to end. 

Second, genuine investment in workforce development, not just reskilling workshops, but the kind of capability transformation that prepares people to work iteratively and critically alongside AI systems. 

And third, treating the human-AI interface itself as a strategic asset, something to be designed with the same rigor as a product or a customer experience, because increasingly, it is the customer experience.

That outcome isn’t guaranteed. It has to be designed.

The right question for leaders – in business, in policy, in education – isn’t how many jobs AI will take. It’s how intelligently, how equitably, and how humanely we redesign the ones that remain.

Vlad Vaiman is a professor at California Lutheran University School of Management and the Chair of the Ventura County Economic Development Association.