The AI training gap in most organizations does not run between people who want to learn and people who refuse to. It runs along the org chart. The people at the top are getting trained, getting more satisfied and getting more comfortable. The people at the beginning of their careers are mostly teaching themselves. And the people with the longest careers carry the most risk when a role disappears.

That is not a technology story. It is a leadership design story.

Three pieces of news from this past week, read together, make the pattern hard to miss.

Key takeaways

  • New ZipRecruiter research finds 83.5% of senior leaders have some AI training from their employer, compared with 45.3% of entry-level workers.
  • Senior leaders are far more likely than entry-level workers to say AI has improved their work satisfaction (76.5% vs. 39.8%).
  • Goldman Sachs research reported by Fortune suggests workers displaced from AI-exposed jobs take longer to find new work, with a wider gap for workers 35 and older.
  • Closing the AI training gap is not a generational initiative. It is a decision about who gets to grow inside the new structure of work.

What’s happening

On October 26, ZipRecruiter Economic Research released Better Work, More to Learn: The 2026 AI Worker Report, based on a survey of more than 1,500 U.S. workers fielded September 8–21, 2026. Most respondents (82.5%) use AI in some form at work, and 52.6% say it has made their work more enjoyable. But the benefits are not landing evenly. 76.5% of senior leaders say AI has improved their work satisfaction, compared with 39.8% of entry-level employees.

The training numbers follow the same line. 83.5% of senior leaders have some AI training from their employer. For individual contributors it is 46.4%. For entry-level workers it is 45.3%, and 54.7% of entry-level workers report learning AI tools on their own. Meanwhile, 49.9% of workers say their employer now has higher productivity expectations. And 25.1% say they plan to leave their job because of how their employer views or uses AI. ZipRecruiter labor economist Nicole Bachaud put it plainly: “the workers without access to formalized training are the ones most at risk of falling behind.”

At the other end of the career curve, Fortune reported on October 30 on Goldman Sachs research on workers displaced from AI-exposed jobs. Using a scoring method Goldman built from occupation exposure, industry adoption and whether the position was eliminated, economist Pierfrancesco Mei found that these workers took about one month longer than other displaced workers to find new jobs, and were roughly 10 percentage points more likely to still be unemployed a year later. For workers 35 and older, the extra time to return to full-time work was about five weeks; for those under 35, about three. Goldman itself says it is “still too early to assess the impact along all of these dimensions,” and the article notes that researchers at Stanford and the Yale Budget Lab do not see widespread, economy-wide AI displacement. The pattern is worth watching, not overstating.

The same day, Fortune profiled a very different approach. Cisco president and chief product officer Jeetu Patel described finding a 27-year-old former startup founder, who joined Cisco through an acquisition, “buried away in the organization doing something unstrategic.” Patel gave him a four-week trial and the AI resources to pursue a plan to refactor the company’s legacy code. Patel also described treating interns as reverse mentors and deliberately building teams that mix experienced operators, systems thinkers and AI-native engineers. In his words: “Either of these groups by themselves aren’t nearly as potent as when all three work together.”

What the AI training gap reveals about the structure of work

In The Next Turn, I describe two changes happening at once. The first is the definition of work: what people believe work is for and what it should give them in return. The second is the structure of work: how work actually flows, now that AI is becoming part of the infrastructure. This week’s news sits right where those two meet.

On the structure side, the baseline has already moved. Half of workers say expectations went up. Most are already using AI. That is the point I make in Chapter 6, “AI Is Not a Tool”: the baseline moves even before leadership has fully defined the rules.

On the definition side, the people entering work are asking whether the exchange makes sense. In Chapter 5, “The Value Exchange (The 3-Year Problem),” I argue that growth has to be visible early, and that retention is a design problem, not a motivation problem. When early-career employees are told to be more productive with AI but are left to figure it out on their own, the exchange gets harder to believe in. A quarter of workers saying they may leave over how their employer handles AI is that exchange showing up in the data.

Here is the part that is easy to miss. The gap is not only about younger workers. The Goldman findings point to experienced workers carrying real risk too. And the ZipRecruiter data shows individual contributors, not just new hires, sitting well below leaders on training access. This is not a story about one generation. It is a story about who the organization is investing in as the work changes.

In Chapter 14, “The Leader as Translator,” I put it this way:

No one is standing outside this transition.

The Next Turn

And a few lines later: “Everyone has to move.” That includes leaders. If senior people are the ones getting the strategic briefing while everyone else gets a login and a higher target, the organization is not moving together. It is fragmenting, and the people furthest from the training are the ones carrying the most uncertainty.

Why it matters for leaders across generations

The Cisco example is useful not because every company should copy it, but because of what it assumes. It assumes that early-career people can contribute something real, that experienced people bring something the work still needs, and that the value comes from putting them in the same room. That is closer to coherence than to accommodation.

Chapter 3, “What Each Generation Believes Work Is,” makes a point I come back to often:

The leadership task now is not to decide which generation is right.

The Next Turn

The task is to build a system where different definitions of work can operate without breaking it. Reverse mentoring is one version of that. Shared training is another. What does not work is letting AI capability pool at the top while the people doing most of the daily work are expected to absorb the change on their own time.

There is also a standards question here. Higher productivity expectations are a signal. The standards underneath them, quality, reliability, sound judgment, have not changed. If people are measured against a faster baseline without being taught how to meet it well, leaders end up rewarding whoever figured out the shortcuts first, not whoever exercises the best judgment. In Chapter 12, “Defining Work Again,” I argue that AI readiness is largely manager readiness, and that leaders need to define AI’s role in the work before employees define it informally. Uneven training is exactly how informal definitions take hold.

Your next turn

Before you approve another AI rollout or raise another target, try these this week:

  1. Map training by level. Who in your organization has had real, hands-on AI training, and who has had an announcement? Look at entry-level, individual contributors and senior leaders separately.
  2. Match expectations to support. For every team where expectations have gone up because of AI, ask what training and time came with that change. If the answer is none, the expectation is ahead of the design.
  3. Pair across career stages. Set up one cross-generational pairing where an early-career employee and an experienced employee work on the same AI-assisted workflow and teach each other: one on the tool, the other on the judgment.
  4. Ask the people teaching themselves. Find out what your self-taught AI users have figured out, and what they are unsure about. Their workarounds are your unofficial AI policy.
  5. Look at your experienced people, too. Ask which long-tenured roles are most exposed to AI and what the plan is for those people before a role changes, not after.

Frequently asked questions

What is the AI training gap at work?

It is the difference between who gets formal AI training from their employer and who does not. In ZipRecruiter’s October 2026 AI Worker Report, 83.5% of senior leaders had some AI training, compared with 46.4% of individual contributors and 45.3% of entry-level workers. Many of the rest are learning AI tools on their own.

Why should managers care if entry-level employees are teaching themselves AI?

Because expectations are rising faster than support. When people learn on their own, they also define on their own what good AI use looks like. That makes quality uneven, makes judgment harder to see, and can weaken the value exchange that keeps early-career employees from leaving.

Is AI training mainly a Gen Z issue?

No. Generational labels are broad reference points, and the data points across career stages. Individual contributors trail leaders on training access, and Goldman Sachs research reported by Fortune suggests older displaced workers face longer job searches. Training is a workforce requirement, not a generational initiative.

Closing

AI is changing the structure of work for everyone at the same time. The question is whether leaders are changing the support at the same time, or only the expectations.

The organizations that handle this well will not be the ones with the most tools. They will be the ones where the person in their first year and the person in their thirtieth year are both learning, both contributing, and both clear on what good work looks like now.

If your AI training budget were mapped onto your org chart today, who would be standing outside the transition? (More on leading through these shifts is on the Next Turn blog.)

Sources