Machine builders have every reason to use artificial intelligence (AI) for repetitive engineering work. It can draft code, generate documentation, assist with design, identify patterns and reduce the time engineers spend on tasks that do not require their highest judgment.

But if companies translate those gains into fewer junior engineers, they may create the next controls-engineering shortage themselves.

The latest Stanford Digital Economy Lab (SDEL) update shows the broader early-career trend worsening. The employment shortfall for workers ages 22 to 25 in highly AI-exposed occupations widened from 15% using the July 2025 data vintage to 19% by June 2026, largely because employers are hiring fewer young workers, according to the SDEL report, “No Widespread Displacement, but the AI Employment Gap for Young Workers Has Widened to 19%.”

Engineering students must learn to audit AI output through open-ended problems that require uncertainty, independent assessment and defensible technical decisions, according to Purdue University’s Dr. Grant Richards.

Machine builders should extend that logic directly into entry-level jobs.

Let AI draft routine programmable logic controller (PLC) code. Then require the junior controls engineer to test failure states, trace dependencies and explain what happens when the sensor signal is wrong rather than absent.

Let AI create a first electrical design. Then have the engineer review maintainability, panel constraints, safety, field conditions and the downstream consequences of a seemingly efficient choice.

Let AI generate commissioning checklists. Then put the junior engineer on the floor when the machine behaves differently from the simulation and make that employee investigate the discrepancy with a senior colleague.

Let AI organize documentation. Then ask the engineer to explain the system architecture well enough to troubleshoot it at 2 am when the original designer is unavailable.

That is the apprenticeship machine builders need in an AI-rich environment.

The industry’s real expertise has never consisted of typing code quickly. It comes from understanding how motion, safety, networks, sensors, drives, software, mechanical systems, operators and production constraints interact when a real machine is expected to work reliably.

Controls engineers need training and guidance as the technology stack broadens. AI raises the stakes because it can make a beginner look productive before that beginner understands the system deeply.

Get your subscription to Control Design’s daily newsletter.

Managers should therefore measure time to independent competence, not merely engineering hours saved.

How soon can a new engineer commission a subsystem without a senior person catching major omissions? When can that employee diagnose whether a failure originates in logic, wiring, networking, mechanical behavior or process conditions? Can the engineer challenge an AI-generated solution because it violates a safety, reliability or maintainability constraint?

If AI accelerates those milestones, machine builders gain leverage from both technology and talent.

If AI simply reduces junior hiring, the industry will continue competing for the same experienced controls engineers while fewer people receive the field experience needed to join that group.

That strategy cannot scale.

The better model is to capture expert reasoning as part of the workflow. When a senior engineer changes an AI-generated design, record why. When commissioning exposes an assumption the digital model missed, turn it into a teaching case. When a junior engineer catches a bad recommendation, review the reasoning with the team.

This preserves institutional knowledge while exposing younger engineers to more exceptions faster.

The widening 15% to 19% early-career employment gap is a signal that companies across AI-exposed fields are narrowing the first rung. Machine builders should resist that pattern because their future depends on people who can take responsibility when automated systems meet physical reality.

The purpose of AI adoption at work should be to eliminate low-value engineering tasks, not the developmental path that creates high-value engineers.