Ask an AI to write a lesson plan and you will have something on the screen in seconds. Ask whether it is any good and the answer gets more interesting. AI can genuinely transform how teachers and homeschooling parents plan - but only if you are clear about what it does well, what it does badly, and where a human has to stay in the loop.
The honest answer is not “yes” or “no.” It is “yes, for some of the job.” Here is how to tell which parts.
Where AI genuinely helps
AI is excellent at the parts of planning that are time-consuming but not where your professional judgement lives. It drafts quickly, so a blank page is never the obstacle. It personalises effortlessly - ask for the same concept framed around soccer, then horses, then Minecraft, and it will not sigh. It differentiates on demand, producing an easier and a harder version of a task in the time it takes to ask. And the best tools now build interactive activities, not just text: sorters, visualisers, self-checking questions.
For a teacher facing thirty different students, or a parent following one child’s shifting obsessions, that speed is not a gimmick. It is the difference between personalisation as a nice idea and personalisation as a daily habit.
Where human judgement is irreplaceable
AI does not know your student. It does not know that one child shuts down when a task looks too long, or that another needs the win of an easy first question before they will attempt anything hard. It cannot read the room, build the relationship, or make the assessment call about whether a learner has truly understood or just performed understanding.
AI can write the lesson. It cannot know the child the lesson is for. That gap is exactly where teaching actually happens.
So the model that works is not “AI replaces the planner.” It is “AI does the heavy drafting; the human steers, judges, and adapts.” You bring the knowledge of the learner; the tool brings the speed.
The real divide: generic filler vs. interest-led learning
Most of the disappointment with AI lesson plans comes from treating it like a vending machine - type “Year 5 fractions worksheet,” accept whatever falls out. That produces exactly the generic material that bored students in the first place. The output is only as specific as the input. A great prompt names the concept, the year level, the student’s actual interests, and the activities you want - and a great tool turns that into something interactive and on-standard, not another page to read.
What this looks like in practice
- Use AI for the draft and the personalisation; reserve the judgement calls - pacing, assessment, tone - for yourself.
- Feed it specifics. A named interest and a clear activity request beat a one-word topic every time.
- Read every lesson as the student would before you use it, and tweak rather than accept.
- Favour tools that produce interactive, curriculum-aligned lessons over ones that just generate text.
That last point is the whole reason Sprout exists. Instead of a generic worksheet, you describe the concept and your student’s interests and get back an interactive, standards-aligned lesson in seconds - AI doing the drafting, you doing the teaching. Used that way, the answer to “can AI write good lesson plans?” is a confident yes.