AI in the Workplace: Why Human Skills Are More Essential Than Ever

For most of the last decade, the loudest advice in the labor market was that technical ability was the priority. Learn to code. Get the specific technical or professional certification. That message seemed to dominate the public conversation for so long.

The market has now corrected, and I believe it has overcorrected in some ways. Open a professional feed and you will likely come across a claim that human, durable skills are the new moat and the last thing AI cannot replace. It has become its own cliché, and it repeats the original error in reverse.

Most institutions teach both halves, mine included, and teach them well. Very few teach them in the same course. Technical skill lives in one course sequence; communication, ethics, and critical thinking live in another, taught by different faculty, assessed on different rubrics, often years apart in a program. Students learn to move between the two rather than to fuse them.

That is our reason behind embedding AI fluency in every course we teach at DeVry, not only the technical ones. The instruction is the same whether the subject is accounting, project management, or engineering technology: Use the tool, then interrogate what it gave you. Students are taught to validate the output against a source, test the logic, identify what the model could not have known, and defend the conclusion as their own work. By the time students graduate, they have done this hundreds of times across a dozen unrelated contexts, and questioning a confident answer has become a reflex.

The reassuring version of this story is that AI redefines work rather than destroying it. The World Economic Forum's projections broadly support it: 170 million roles created against 92 million displaced by 2030, resulting in a net increase of 78 million jobs, with roughly 39% of core skills changing along the way.

I would, however, be careful with the word net. The person displaced from customer operations in Ohio is not the person hired into an AI governance role in Austin, and the distance between those two facts is measured in years, money, and time that most working adults cannot take away from a job and a family.

Our own research at DeVry shows where that distance turns into a barrier. Nearly 9 in 10 employers say they offer company-paid upskilling. By their own estimate, only about half of their workers use it. Something is failing between the benefit being offered and the adult with a full-time job, a commute, and children who need care and attention. Two related findings sharpen it further: 66% of employers say they are concerned that their workforce is not keeping pace with technology, and 59% say they see no reason to hire someone who does not already bring a skills advantage. The skills employers say they cannot find are increasingly the skills they expect candidates to arrive with, fully formed, at their own expense.

Here is a phrase I would like to retire, and I say this having used it myself: AI-proof. Nothing is AI-proof. Any skill defined by its resistance to automation is a skill defined by a shrinking perimeter.

AI literacy deserves the same scrutiny. Literacy is the floor. Knowing how to prompt a model is roughly what knowing how to use a search engine was in 2004; it was briefly a differentiator and then a condition of employment.

Our research puts a finer point on it: Workers place enormous value on AI skills while reporting that they do not know how to use AI tools well. That gap between valuing a tool and wielding it with judgment and critical thinking is the whole game, and another simple tutorial will not effectively close it. What is scarce, and getting scarcer, is fluency: knowing what to ask, recognizing when a confident answer is wrong, and remaining accountable for the decision after the machine has spoken. These systems generate plausible output at extraordinary volume. Plausible is not correct, and someone has to be able to tell the difference.

This is the argument I made when DeVry was featured in The Skills of Tomorrow, a documentary segment produced by Acumen Media for the Global Sustainable Trade Initiative series, published on CNBC. The film examines how organizations across industries are adapting to technological change, and what I wanted to convey in our segment was less about our programs than about a design choice: a dual-competency model, in which technical AI fluency is developed alongside discernment, ethical reasoning, and the capacity to think critically.

Reskilling is usually framed as individual responsibility: a matter of initiative, of staying current.

Return to the finding that 9 in 10 employers offer upskilling and roughly half of workers take it up. The easy reading is that people are not motivated. I have spent more than 2 decades around this population, and I do not believe that for a moment. The learners DeVry serves are enrolling in programs while holding down jobs and juggling several competing priorities. What they are short of is not will. It is capacity: the hour that is not already claimed by a shift, a commute, a child, or a second job.

The deeper problem is that fluency compounds where it already exists. People whose work puts them near these tools use them daily, get better, and grow more confident about what the technology will do for their careers. People whose jobs keep them at a distance get none of that, and the distance widens over time. This is why I resist the idea that AI is an inherently democratizing technology. It democratizes capability only for those already close enough to touch it. The World Economic Forum makes the same point when it notes that how these roles evolve carries direct consequences for economic mobility.

That is not a talent problem. It is a distribution problem, and distribution problems have solutions.

DeVry has spent 95 years serving the people furthest from that door. What we teach has changed many times. The mission has not changed: to close society's opportunity gap and prepare learners to thrive in careers shaped by continuous technological change.

The question is not whether our graduates can use AI in the workplace. Within a few years, everyone will. What will distinguish them are the durable, human-skills, and durable skills can be taught. That is why I am optimistic about the decade ahead.