What young people learn when AI writes the code — treating the agent as something to direct and check, not an answer machine
The National Science Foundation awarded roughly $1.32 million to the University of California-Irvine to develop the Agentic Coding Studio, a platform where young people work with an AI coding agent to create games, stories, simulations and applications. Rather than treating AI as an answer machine, the design keeps youth responsible for what they make.
Grant overview (primary data)
- Award amount$1,315,321
- RecipientUniversity of California-Irvine (California)
- ProgramAISL
- Period2027-01-01 〜 2029-12-31
- FunderU.S. National Science Foundation (NSF) / NSF
Key points
- Development of the Agentic Coding Studio, where young people work with an AI coding agent to create games, stories, simulations and applications ($1,315,321; 2027-01-01 to 2029-12-31).
- The institution is the University of California-Irvine, under AISL, which covers learning outside classroom instruction.
- The premise is that as AI takes on routine code production, people must define goals, break problems into parts, judge solutions, test results and revise.
- Rather than treating AI as an answer machine, the design keeps youth responsible for what they make.
- Of the 120 NSF awards this site holds as of 2026-09-02, 4 fall under AISL; the median obligated amount is $1,955,351 and the median duration three years.
1When the writing shrinks, what work remains
Learning to program has long centered on absorbing syntax and becoming able to write code by hand. As AI agents take on routine code production, that center moves. The description names what remains on the human side quite concretely: defining goals, breaking problems into parts, judging proposed solutions, testing results and revising the work. Deciding and verifying come forward; writing recedes.
2Not treating it as an answer machine
The contrast matters as instructional design. Handed the same AI, a student who delegates entirely and one who keeps judgment on their own side come away with different things. That the artifacts are described as personally meaningful also works as a device that makes handing judgment to someone else harder.
3The setting outside school
The program is AISL, Advancing Informal STEM Learning, covering science and technology learning outside classroom instruction. Of the 120 NSF awards this site holds as of 2026-09-02, 4 fall under AISL. There are 70 distinct programs, most directed at research infrastructure or workforce development, but awards designing the learning setting itself hold a steady share.
Learning outside school runs on the learner own interest rather than grades or credit, which makes it a different sort of place to test how people work with AI.
4Duration and scale
The obligated amount is $1,315,321 over three years, from 2027-01-01 to 2029-12-31. Across the records this site holds as of 2026-09-02, obligated amounts have a median of $1,955,351 and durations a median of three years, so the duration is standard and the amount somewhat below the median. That the start is set about four months after the award date also reflects a project that has to prepare materials and a setting. Amounts are the obligated amount as of the check date and may change.
Why it matters
The view that required skill shifts from writing toward deciding and verifying as coding agents spread carries into workplace training as much as education. That delegating entirely and keeping judgment produce different learning is worth weighing when setting internal adoption policy.
FAQ
If AI writes the code, is there a reason to learn programming?
What is AISL?
Sources (primary)
Source: NSF Award Search (U.S. National Science Foundation, public domain). Amounts are the obligated amount. For privacy, we do not handle principal investigator names.
- NSF Award (original, official)
- NSF Award ID: 2628148