1,600 Teachers Just Got AI Mentors. Access Was Never the Point.
This week in Jonesboro, Georgia, a room full of Clayton County teachers sat down with OpenAI mentors and rebuilt their lesson plans, their parent emails, and their Monday morning paperwork with AI at the table. Seven more cities get the same treatment between now and September. It looks like a product tour. Read the fine print and you find an admission.
OpenAI Academy and the Walton Family Foundation are running what they call an AI Skills Jam for K12 educators, an in person summer series reaching more than 1,600 teachers, administrators, and district leaders across the country. Jonesboro and Fairfax kicked it off this week. Orlando, Chicago, San Bernardino, Phoenix, Salt Lake City, and Las Vegas follow through early September.
Here is the admission buried inside the announcement: the company that gives away ChatGPT for free just spent real money to put human mentors in physical rooms with teachers. Access did not close the gap. Practice does.
The Tool Was Never the Hard Part
Every teacher in those rooms already had access to AI. Most have had it for three years. Districts bought licenses, IT departments held rollout meetings, and usage still stalled at the copy and paste stage. The workshops target the actual work: lesson planning, parent and staff communication, administrative tasks, the daily grind that eats a teacher’s evening.
Now, you might be thinking: this is a marketing play by a company that wants schools locked into its platform. Partly, sure. But watch what they are spending money on. Not ads. Not free credits. Mentors, chairs, and repetition. When the biggest AI company on the planet decides the bottleneck is practice rather than access, that tells you where the real gap lives.
I call that distance the Practice Gap: the space between having an AI tool on your laptop and having the judgment to make it carry real work. Licenses do not close it. Policies do not close it. We covered in Your AI Policy Is Not AI Literacy how districts confuse rules with skills. This program is what closing the Practice Gap actually looks like: sit down, bring your real lesson, build it with someone who has done it before.
The Number That Sells the Whole Program
The announcement leans on one statistic, and it is a good one.
Run the math on your own week. Six weeks a year is not a rounding error. That is grading returned faster, parent emails that stop piling up, and a Sunday evening that belongs to your family again. The teachers pulling those hours back are not doing anything exotic. They use AI weekly, on real tasks, with enough skill to trust the output after they check it.
The trap hiding in that stat is the word weekly. The savings show up for teachers who practice consistently, then compound. The teacher who opened ChatGPT twice in October and gave up gets nothing. Worse than nothing, actually: they carry the same workload plus the quiet suspicion that everyone else figured out something they missed. That suspicion is the real cost of the Practice Gap, and it is why one workshop, or one summer, will not finish the job on its own.
Access to AI is just the starting point, the real opportunity is building the agency to use these tools thoughtfully.
Leah Belsky, VP of Education, OpenAI
Agency is the right word, and it is the one most AI training skips. That insight matters on its own. But the next question is the one that decides whether this program changes anything for you.
What Happens to the Other Three Million Teachers
Like what you’re reading? Get insights like this delivered daily.
Eight cities. Sixteen hundred educators. The United States has over three million teachers. Even a sold out tour reaches a fraction of one percent of the profession, and participants get follow up through OpenAI Academy, a free online library of guidance and best practice. Useful, yet a library is access again. The in person mentor, the real task, the immediate feedback: that is the part that worked, and that is the part that does not scale on OpenAI’s budget.
It scales on yours. The structure of a Skills Jam is not proprietary. It is a room, a real task, a person slightly ahead of you, and repetition. Any department meeting can run one. Any grade level team can run one. If you are anything like most educators reading this, your objection is already loaded: who has the time to organize that? The answer comes from the same Gallup number. The teachers in your building already saving six weeks a year have the time. Borrow one of their planning periods and you both come out ahead.
Here is what a homegrown version looks like, and notice it is the same shape as the build it, do not announce it approach that worked for faculty micro credentials:
- Bring one real task. Not a demo prompt. The actual unit plan, the actual newsletter, the actual rubric due this week.
- Pair up around the strongest user in the room. The mentor does not need to be an expert. They need to be twenty hours ahead of everyone else.
- Finish the task before the session ends. The win has to be real and immediate. Walking out with shipped work builds the agency Belsky is describing.
- Repeat on a schedule. One session builds a memory. A monthly cadence builds a skill. This is the compounding we teach inside SeedStacking, where small consistent reps beat occasional heroics every time.
Want a community that runs on exactly this rhythm? Join the free Harvest Kernel Learning Community
Want this as a worksheet you can run with your team? The AI Skills Jam Starter Kit walks all four phases with checklists and an hours saved tracker.
The Mistake That Will Waste This Moment
Most coverage of this program will score it as an adoption story: OpenAI woos schools, schools sign up, line goes up. That framing repeats the mistake that burned the last three years of AI rollouts, measuring seats instead of skills. A district that buys licenses and skips practice ends up with the pattern we broke down in They Traded People for Tokens: spend on the tool, skip the people, wonder where the returns went.
The reason this program matters is that it prices the missing ingredient. When OpenAI itself invests in mentors and folding chairs, the message to every principal, department chair, and professional development coordinator is plain: budget for practice, not just platforms. Teachers who get structured, hands on repetition with real tasks reclaim measurable time. Teachers who get a login and a memo do not.
The Harvest Kernel Takeaway
OpenAI just confirmed what the license invoices never will: access is the starting line, practice is the race. You do not need to wait for a mentor tour to reach your city. One real task, one person twenty hours ahead of you, one finished win, repeated monthly. That is how six weeks a year comes back to you.
Ready to go beyond reading and start building AI fluency?
Join the free Harvest Kernel community for practical guidance, fresh ideas, and tools that help you make AI useful in real life.
Sources
1 OpenAI, “Helping K12 educators build practical AI skills,” July 8, 2026. openai.com/index/k-12-educators-practical-skills
2 Walton Family Foundation and Gallup, “Teaching for Tomorrow: Closing the Expectations Gap,” cited in the OpenAI Academy announcement above.
