Choose according to the learning problem, not the product label. For this guide, a coding tutor develops programming foundations, an AI mentor develops AI understanding and projects, and AI tutoring software is a tool rather than the human service being purchased.
Explore the comparison in this briefing
Your child’s immediate need
Cannot comfortably read or modify a simple program
- Suggested starting route
- Coding tutor
- Evidence to request
- Clear explanations, small exercises and independent modifications
Compare each option against the same criteria. The complete table and qualifications remain in the briefing.
Stanford’s August 20, 2026 briefing distinguishes human-led tutoring supported by AI from instruction delivered directly by AI software. It cautions against treating these models as interchangeable or assuming that software inherits the evidence base for high-impact human tutoring. (nssa.stanford.edu)
| Your child’s immediate need | Suggested starting route | Evidence to request |
|---|---|---|
| Cannot comfortably read or modify a simple program | Coding tutor | Clear explanations, small exercises and independent modifications |
| Wants to understand AI’s capabilities and limitations | AI-literacy teaching | Activities involving evidence, mistakes, bias and responsible use |
| Can code and has a defined application idea | Project-focused AI mentor | Relevant technical experience and a realistic learner-owned scope |
| Wants additional practice between lessons | Age-eligible educational software | Appropriate oversight and a clear role within the learning plan |
| Needs a finished application by a deadline | Clarify education versus development services | An explicit division of student, mentor and contractor work |
Buying distinction: if the instructor mainly delivers a product, evaluate it as development work. Do not count all of that work as the child’s learning.