What AI genuinely wins at
- Availability. 2am, the night before, for the ninth time. A human tutor is asleep, and rightly.
- Infinite patience. You can ask the same question six ways without embarrassment — and embarrassment is a genuine barrier that stops people asking.
- Cost. Not close.
- Unlimited practice questions, generated on demand, in your exam's format.
- Immediate feedback, which matters more than most people realise — the value of feedback decays with delay.
What a human still wins at, decisively
Diagnosing what you're avoiding. A good tutor watches your face, notices the hesitation, and pushes exactly where you don't want to go. An AI answers the question you asked — and the question you asked is, very often, not the question you needed.
Accountability. A person expecting you on Thursday is a commitment device. A chatbot is not, and no amount of notification design changes that.
Judgment about what matters. An experienced teacher knows which topics carry the marks and which are examiner obsessions. An AI without your syllabus does not — which is exactly why grounding it in your actual course material closes some of this gap.
Knowing when you're bluffing. The single biggest one. You cannot fool a tutor who asks a follow-up. You can absolutely fool yourself with an AI that accepts your framing.