Yesterday morning, I attended the Citizens Budget Commission’s panel discussing “How AI Will Reshape New York’s Economy and Government.” On the panel, Council Member Virginia Maloney, as Chair of the Economic Development Committee, largely spoke about attracting AI companies to the city and AI skills training for workers and business owners. She also spoke to something that’s at the center of the AI and education debate: judgment.
How do we teach judgment and discernment when those skills are usually gained through trial and error at work? Can we rethink the education system to prioritize those skills as we lose the entry-level jobs that are supposed to teach them?
Let’s dive in.
Jared Giles
Founder, SquareOne
[email protected]
Judgment is a Learned Skill
In the first issue of The SQ1 Pivot, I argued that our talent pipelines should connect directly to the massive economic infrastructure projects that are set to bring tens of thousands of high-wage jobs to our neighborhoods. AI raises the stakes on that argument, because the skill many of those jobs will demand most is the one young people are losing the chance to build.
Fewer young people are getting in the door: Employment for 22 to 25 year olds in AI-exposed occupations is now about 19% below where it would be if it kept pace with less-exposed occupations.
The jobs that stay require more: AI-exposed entry-level roles are 7 times more likely to require senior skills like judgment and leadership.
A junior employee’s ability to practice judgment used to come from their first job or set of jobs. Creating a space where a young person can work and be coached by those around them while making mistakes and learning from those mistakes is the benefit of the entry-level job. But if employers are hiring fewer young people and expecting senior-level judgment on day one, students need those experiences before they graduate high school.
If AI is taking the work that taught young people what good looks like, high school has to become the place they learn it.
Career-connected learning is how students get those experiences. In Connecticut, young men who attended a technical high school earned about 32% more in their first years of work than similar applicants who missed the admissions cutoff, and roughly a fifth of that gain traced back to work experience tied to what they studied. Delaware’s 2026 pathways study found workplace experience mattered most in career pathway programs, and students who did internships or apprenticeships were more likely to persist in jobs in the same fields. And in New York City, MDRC’s study of P-TECH schools found students were 38% more likely to have an internship during high school and 5% more likely to finish an associate’s degree within seven years. The city’s largest economic investments are where the education-to-career connection should happen. SPARC Kips Bay will bring 3,100 permanent life sciences jobs to a campus planned with space for NYC Public Schools and CUNY, and JFK’s $19 billion redevelopment is creating 9,600 direct jobs. Students should be learning inside of those spaces.
Students also need to practice judgment on the tools themselves, and that should start no later than 9th grade. Because NYC’s AI policy pauses student-facing AI through 8th grade, 9th grade becomes the first year students can use it in school. Right now that year only includes two 45-minute AI literacy modules, with more leeway for students in career and technical education (CTE) programs. At the same time, 54% of teens are already using chatbots for schoolwork. AI belongs in at least 9th grade as a technical skill, taught alongside the certifications and core skills their CTE programs prioritize.
With access to the right tools and opportunities, a student at SPARC could use AI to draft discharge instructions for patients, then sit with a supervisor who walks through where the AI messed up and why it matters. Today that happens after high school, and often after college. It could happen in 11th grade. A student who has done that kind of work hundreds of times by graduation walks into their first job with the judgment and discernment to succeed.
National Spotlight: Gwinnett County, Georgia
How do you teach students to use AI with judgment instead of banning it? Look at Gwinnett County.
The Problem: When Gwinnett started building its AI programs in 2019, the district realized it couldn’t limit AI learning to high school, because students’ perceptions, beliefs, and habits around technology form when they’re very young.
The Intervention: The 177,000-student district opened what some call the nation’s first AI high school. At Seckinger High, every student focuses on becoming a responsible and ethical user, decision-maker, and problem-solver with AI, and those who want to go deeper can take a three-course AI pathway. Seckinger graduated its first four-year cohort in 2025, and the pathway is spreading: Paul Duke STEM High opens its own AI pathway this year, and any student in the district can access it at Maxwell High School of Technology. All of this sits inside a system that prioritizes career-connected learning, and some of the AI teachers came from industry as career changers.
The Takeaway: Gwinnett treats AI judgment as something taught in the classroom and doesn’t rely on students picking up the skill on their own. Since the AI director is aware that there aren’t many entry-level AI jobs right now, the goal is to prepare students for college and career.
ICYMI
[WATCH] Citizens Budget Commission: Two panel discussions covering how AI will reshape New York’s economy and how we can leverage AI in government.
[REPORT] Stanford Digital Economy Lab: Data showing how the employment gap for young workers is widening, particularly in jobs that rely on “codified knowledge.”
[REPORT] PwC: A report detailing how entry-level roles are being rewritten to demand senior judgment.
Stay connected and explore our latest work at SquareOne.
