By Gleb Tsipursky, PhD
New Britain already has two pieces of the workforce model the AI era needs. Central Connecticut State University’s AI Corridor is built around applied research, hands-on learning and industry-connected projects. This spring, Connecticut also highlighted newly certified apprenticeship coursework at Lincoln Technical Institute in New Britain, where students combine rigorous instruction with the employer-led structure of registered apprenticeship.
The next step is to put those ideas directly into AI-enabled entry-level jobs.
Stanford’s August employment update finds employment among U.S. workers ages 22–25 in highly AI-exposed occupations about 19% below the path implied by similarly aged workers in less-exposed occupations. The adjustment is concentrated in reduced hiring. Experienced workers show no comparable gap.
Connecticut is simultaneously pushing AI skills deeper into the workforce. The state’s Tech Talent Accelerator reports that nearly 11,000 Connecticut job postings have required AI skills since August 2024, up 40% from the prior year. That demand creates an opportunity for New Britain to shape how AI skills translate into careers.
Employers should create an AI apprenticeship layer for junior roles. AI can draft the routine memo, assemble the first spreadsheet, summarize the intake file or produce the first pass at a research task. The junior employee should then verify the evidence, identify missing information, test edge cases, explain what the system got wrong, and make a bounded recommendation under review.
That sequence matters because professional judgment grows through repeated practice. An entry-level employee who only accepts AI output becomes faster at operating software without becoming much better at the job. An employee who must inspect, challenge and improve AI output develops the pattern recognition that experienced workers rely on.
New Britain employers can make this concrete with three operating rules.
First, every AI-assisted junior role should have a named experienced reviewer and protected review time. Mentor time should be treated as productive capacity rather than overhead.
Second, supervisors should deliberately move beginners through progressively harder work. Start with verification, then exceptions, then recommendations, then supervised decisions. AI should shorten the preparation stage so employees can reach those higher-value tasks sooner.
Third, measure time to independent competence alongside hours saved, output and cost. Track error detection, rework and the number of tasks a junior employee can complete without escalation. Those measures reveal whether AI is creating stronger talent or merely reducing visible labor.
New Britain has long understood workforce development as something built through institutions, employers and practical experience. AI should strengthen that tradition. The city can become more productive while giving young workers more chances to become the people who know what to do when the easy answer is wrong.
========================================
Gleb Tsipursky, PhD, a behavioral scientist, CEO of Disaster Avoidance Experts, and author of The Psychology of AI Adoption at Work: From Resistance to Results (Georgetown University Press, 2026). https://disasteravoidanceexperts.com/aibook
