Home › Hard problems › Society and trust

Society and trust

An AI tutor that measurably raises learning

Solo: possibleOne person can test a narrow slice of it.

studentAI tutorhelp?"the answer is 42"Feels helpfulhomework done,student happyTest without itgain: rarelyshown in trialshow it feelsmeasured gainsome students learnto lean on it instead
Many AI tutors feel helpful, but few show test-score gains in controlled trials, and some teach students to lean on them.

Why it is still open

Many tutors feel helpful, but few show test-score gains in controlled trials, and some make students lean on them.

An experiment one person can run in 8 weeks

Run a 6-week trial with 2 groups of students on one topic, comparing a hints-only tutor with one that gives answers. Measure scores before and after.

Verdict: Solo: possible

Needs willing students and careful design.

The solo test behind the verdict

  1. Can it be measured? There is a number that shows you did better, such as an error rate, a catch rate or a cost per answer.
  2. Is it cheap to try? It runs on a laptop, free open models and open data, or under about ₹50,000 of rented GPU time.
  3. Is there a public yardstick? A benchmark or dataset others use, so your result can be compared and believed.
  4. Would a small win get noticed? A lab, a funder or a customer would care about a 10% improvement, not only a full solution.

Four yeses means Solo: strong. Two or three means Solo: possible. Fewer means Follow, don't build.

PreviousProving where content came fromNextFact-checking at the speed of rumour

All 79 hard problems · Working on this, or know of a result we missed? basant@cleandataengineering.ai