Strategy, Artificial Intelligence & Entry Level Hires

Humility vs. Delusion

Source: Roger L. Martin, 2025

As I promised at the inception of Year V of the Playing to Win/Practitioner Insights (PTW/PI) series, I am doing co-authored pieces this year. I led off the year with the first one (Beyond Network Economics) and this is the second. For this article about Artificial Intelligence (AI) and entry level hiring, I am joined by an AI company founder — Ahmad Zaidi, Co-Founder and CEO of TransforML — and a senior executive of the world’s biggest private sector employer — Gui Loureiro, Regional CEO Walmart Canada, Central America, Chile and Mexico and co-author of Reinventing the Leader.The piece is called Strategy, Artificial Intelligence & Entry Level Hires: Humility vs. Delusion. And as always, you can find all the previous PTW/PI here.

The Provocation

A few weeks ago, Dario Amodei, one of the leading thinkers in the AI space and the CEO of Anthropic, a leading AI player and home of Claude, made the bold claim: AI could wipe out half of entry level jobs in the next 5 years.

This coincides with claims from CEOs of leading technology companies like Microsoft and Salesforce about AI writing a significant proportion of code, and the New York Times concluding that millions of young people are graduating to ‘look for work in industries that have little use for their skills, view them as expensive and expendable, and are rapidly phasing out their jobs in favor of artificial intelligence.’ And according to SignalFire’s State of Talent Report 2025, it is already well underway. Tech companies indicated that they had already reduced their 2024 entry-level hiring by 25% versus 2023. The reduction in 2025 may turn out to be even steeper!

Intervening in Complex Adaptive Systems

We believe that companies are engaging in well-meaning efforts to make a single apparently meritorious intervention into a complex adaptive system (CAS) — which is defined as (based on early work by Dooley) “a group of semi-autonomous agents who interact in interdependent ways to produce system-wide patterns, such that those patterns then influence behavior of the agents.” The Amazon jungle, a large city, and the Internet are all CAS — as is any company.

And this is by no means the first time well-meaning people have made purposeful interventions into a CAS. In the early 20th century, policy makers concerned about its burgeoning rodent population added the South American Caracara falcon into the remote and pristine Easter Island ecosystem — the most remote inhabited island in the world. It seemed like a good idea at the time, except that it turns out that Caracara falcons aren’t rodent hunters. But they love feasting on the eggs of other bird species — especially those that they can intimidate as falcons are inclined to do. They didn’t do the job their addition was designed to do, while they wiped out the entire seabird population of the island.

At approximately the same time, US Interior Department officials sanctioned the subtraction of wolves from Yellowstone Park to prevent them from killing livestock in surrounding farms. Their full elimination by 1926 had a famously negative ripple effect on the populations of innumerable species in the park, though grew elk population mammothly, which ruined the vegetation ecosystem and the many animals that depended on it. That process and the wonderful reversal of all the negative consequences with the reintroduction of the wolf beginning in 1995 is chronicled in this short video.

The AI/Entry Level Hiring Intervention

Based on the reports, the modern business world is in the process of simultaneously making an addition — use of AI to replace managerial tasks — and subtraction — reduction/elimination of entry level hiring — intervention to the CAS of the modern company. Without question, a modern company is a CAS. It is complex — the relationships between the heterogeneous actors within and around a company are difficult to map. It adapts — when anything impinges on it, there is adaptation. And it operates as a system. The business world is implicitly assuming that the AI-for-entry-level-hiring substitution won’t damage the company CAS in ways the addition of Caracara and subtraction of wolves most certainly did.

We believe there are at least two critically important assumptions that the architects of this move are making, with about as little consideration of the working of the CAS as the architects gave in the Caracara addition and wolf subtraction decisions:

· We will develop the requisite number of senior people without nearly as many entry-level hires, and

· The culture will not change for the worse when there are many fewer junior people around.

Developing the Requisite Number of Senior People

The current company CAS turns entry level hires into senior executives at a rate that is sufficient to populate the senior levels of the company on a continuous basis. Is that system perfect? Hardly. Often, a company must recruit from outside to fill senior positions. But overall, the larger business system is in rough equilibrium. There are enough entry level people in the overall business ecosystem that senior positions have a chance to be filled — whether internally or externally or with a combination of both.

It is assumed by AI enthusiasts, and apparently many senior executives, that unlike senior roles, entry level roles are sufficiently rote and algorithmic that they can be replaced by AI. Let’s assume that is true.

But if we eliminate or dramatically scale back entry level employees, how will the CAS of the company produce as many senior executives as is does now? One possibility is that it won’t do so, but a company may think that it can hire senior people from outside rather than develop them from within to make up the shortfall. But this is where the CAS comes back to bite. Overall, it is a zero-sum game in which each company facing this shortfall is as likely to have net senior people poached as it is to add incremental senior people.

Alternatively, the belief could be that learning and development could be made to happen so much more efficiently that a smaller number of entry level hires will still produce the requisite number of senior executives. Sadly, there is no data or even argument put forward that this is the case. Or perhaps with less competition among a smaller number of entry level hires for opportunities for advancement will make them more likely to advance more quickly. But it is equally probable that less intense competition may well slow their progress.

The fact is that we have no idea how the CAS of turning entry level hires into senior executives will change with a dramatic reduction in entry level hiring, just like the Caracara and wolf architects had zero idea of how those would impact the respective systems.

Maintaining the Culture

The company CAS currently has lots of junior entry level people in it. Junior people tend to be more excitable, explorative, louder, more active, and often more social. The implicit assumption is that if we remove a lot of them from the CAS, there won’t be change in the culture.

One of us (Roger) had an experience of the opposite sort when he served as dean of a business school. He inherited an aging faculty because the school had been struggling and had difficulty hiring junior faculty. Of the 36 full-time tenure stream faculty, only two were junior non-tenured faculty. An element of his strategy was to more than triple the full-time tenure stream faculty during his fifteen years as dean and that meant hiring a lot of junior faculty. By the time the school reached 120 faculty, the ratio for untenured to tenured was about 50–50, a thirty-fold increase in untenured faculty.

The unexpected cultural benefit was that the senior professors loved having lots of juniors around — their energy was infectious. Many of the original senior professors revivified their careers. The level of excitement in the school rose dramatically. Students loved it.

Within a few years, a dramatic reduction in entry level hiring will compound dramatically alter the age and tenure profile of companies. They will become older and longer tenured. As the earlier examples illustrate, you can never know for certain how a single meaningful change will alter any CAS. Maybe having an older, longer tenured workforce will improve company culture. But it is at least equally possible that the reduction of younger, newer blood in the mix from the current level will damage culture in ways that are hard to forecast. And with any CAS, forecasting the nature/direction of the change (or lack thereof) is a mug’s game.

Practitioner Insights

This is a fundamental strategy choice for business leaders. The opposite — not cutting entry level hiring — is not stupid on its face. And it is an important choice that will have long term consequences. But for many leaders, AI is such an exciting (and well-hyped) new tool that they feel compelled to use it maximally. And their target is what they see as an unnecessarily high numbers of junior workers, whose jobs they believe can be performed cheaper and maybe even better by AI.

It is not unlike Chinese political leaders who came up with a powerful tool — the one-child policy — to tackle what they saw as a population crisis. They kept the tool in place for 35 years. It did the job. Chinese population slowed dramatically. But now China is facing irreversible population aging and decline. Its demographic profile is becoming dangerously aged, and its population will shrink from the current 1.4 billion to between 525 and 800 million (depending on whose estimate you believe) by the end of the century. That level of shrinkage is unprecedented in human history. It may turn out to be fine, but it may be a disaster. Regardless, at this point, there is literally nothing China can do about it.

Demographics in a giant country is a CAS. It turns out that when you subtract a one-week-old, forty years later you have one less forty-year-old! Taking one singular dramatic action while assuming that the CAS will continue to operate the same, with the exception for the one thing that you wish to change, is arrogant and delusional. It is not thoughtful strategy choice.

Everything that the world has learned about CAS — from Easter Island birds to Yellowstone wolves to Chinese population dynamics — points to the need to treat CAS with humility and restraint. Like all important strategy choices, the impact of the choice is long-lasting. A sudden 25% year-over-year cut to entry level hiring is neither humble nor restrained. It is arrogant and delusional.

The trickiest thing about sharply changing a CAS is that it is impossible to know in advance what the unexpected and adverse impacts are going to be. The only thing that is certain is that there will be unexpected and adverse impacts.

Highly intelligent US public health officials decided to intervene in the CAS of K-12 intellectual and knowledge development by shuttering schools for the better part of two years to fight the spread of COVID. Sadly, this turned out to be a Caracara story. There is zero evidence that the intervention reduced or slowed the spread of COVID. However, there is a mountain of evidence that the educational and mental health progress of American kids was severely damaged — and probably permanently so. The lack of humility and restraint of those intelligent US public health officials looks dreadful now.

At least some business leaders appear to be headed down a similar path now. It is time to pump the brakes and show some humility and restraint. Tweak the CAS of your company, don’t body-slam it!

******

As a reminder, I am doing a PTW/PI podcast series with friend Tiffani Bova. The tenth in the series will be on LinkedIn here on Wednesday, July 16th at 12 noon EST and 9am PST. I look forward to seeing you there.