Scorecard

How should you choose an AI mentor for your child?

Check safeguarding and tool eligibility first, then score the named mentor against a 100-point rubric that gives the most weight to demonstrable learning and teaching quality. Do not allow technical prestige, a polished application or convenient scheduling to compensate for an unresolved safety issue.

This scorecard evaluates human-led, one-to-one AI tuition and mentorship, not standalone chatbot subscriptions. Use it to compare private AI tutors for teens, assess a proposed project plan, structure a trial or decide whether to renew.

The weights, assessment activities and decision thresholds below are the desk’s proposed purchasing methodology, not a validated educational instrument or safeguarding certification. No supplier has been scored through observed lessons for this publication.

The decision in four steps

  1. Define the purchase: learner age, starting skills, interests, intended project, available time and maximum commitment.
  2. Clear every applicable gate: people, communication, accounts, data and project permissions.
  3. Collect evidence: documents first, then a supervised teaching trial and a short learning check at a later session.
  4. Decide: compare supported scores, unresolved evidence and the child’s willingness to continue—not the presentation alone.

The parent approves the purchase and permissions. The assistant coordinates evidence and logistics. The learner can decline the match. These are recommended household responsibilities, not assumptions about any particular family.

Sources and context

This page sets out the desk’s purchasing framework. What are the limits of this research? → Sources →