The basic idea
Learners differ in what they already know, how fast they pick up new material and where they get stuck. A teacher with a large class cannot always tailor every task. Software that adjusts to the individual is an attempt to close that gap. The general approach is often called adaptive learning, and it usually works by watching how a learner responds and then changing what comes next.
The wider case for tailoring study to the person is made in the discussion of personalized learning. This article stays with the software side: how the adjustment happens and what to be careful about.
What gets adjusted
Adaptive systems typically change one or more of the following:
- Difficulty. After several correct answers, the questions get harder. After repeated mistakes, they ease off or offer a hint.
- Sequence. The order of topics may change, skipping material a learner has shown they know or returning to a prerequisite they have not mastered.
- Amount of practice. A topic with frequent errors might get more practice, while one with consistent success gets less.
- Timing. Some systems schedule reviews at spaced intervals, in line with ideas from spaced repetition.
- Format. Explanations may be offered as text, worked examples or practice problems, depending on what the system predicts will help.
How the software decides
At the simplest level, rules do the work: for example, move on after a certain number of correct answers. More elaborate systems build a statistical model of what the learner probably knows, updating it with each response. Machine learning can be used to find patterns across many learners, such as which mistakes often point to which misunderstanding.
The important point for a non-specialist is that the system only knows what it can measure. Typically that means answers, time taken and sometimes clicks. It does not see whether a learner is tired, anxious, distracted or bored. A system can therefore misjudge someone, for instance by treating a careless slip as a gap in understanding.
Possible benefits
- Practice that feels neither too easy nor too hard, which may help keep attention
- Less time spent repeating material already understood
- Quicker feedback than waiting for marked work
- Information for teachers about which topics a group finds hard
- A way for self-directed learners to get structure without a class
These are plausible advantages, not guarantees. How much any product helps depends on the quality of its questions, its design and how it is used alongside other teaching. Independent evidence is often limited, and claims from the companies selling a product should be read with that in mind.
Limits of adaptive practice
Narrow measurement. Multiple-choice and short-answer items are easy to score automatically. Skills such as discussion, creative writing or building something are harder to adapt to, so adaptive tools tend to be strongest for well-defined skills like arithmetic facts or vocabulary.
Fixed paths. If a system keeps steering a learner toward easier material, it can quietly lower expectations. Learners need chances to attempt challenging work, including with support.
Loss of the human element. Encouragement, curiosity and explanation from a person matter. Adaptive software is best seen as a supplement to teaching and to the broader learning plan, not a replacement.
Gaming the system. Some learners learn to click through hints or guess to get the next screen, which can make the data look better than the learning is.
Privacy questions
Personalization depends on data, and that raises questions worth asking before a learner begins:
- What information is collected, and is it more than needed to run the tool?
- Who can see it, and is it shared with other companies?
- How long is it kept, and can it be deleted?
- Can a parent or adult learner see the record the system holds?
- What happens to the data if the service closes or changes owners?
Schools often have their own review process for tools, and families can reasonably ask about it. For children especially, laws and school policies about data differ by place, so check what applies where you live. Good digital citizenship habits, such as sharing the minimum and understanding settings, apply here too.
Fairness questions
Systems learn from data about past learners. If that data underrepresents certain groups, the system may work less well for them, for example by misreading answers written in a different dialect or by assuming background knowledge that not everyone has. There is also a risk of labeling: a learner placed on a lower track may stay there longer than is helpful.
Other fairness issues include unequal access to devices and reliable internet, and whether the tool supports learners with disabilities. A fair system lets people question or override its decisions, and keeps a teacher involved in important choices.
Using these tools sensibly
If you are considering an adaptive tool for yourself or a child, a few steps are reasonable:
- Decide what skill you want to improve and whether this kind of practice fits it.
- Try a free or trial version first, if available, and watch whether the learner stays engaged.
- Check the privacy information in plain terms.
- Keep an eye on whether the difficulty feels right, and adjust or stop if it does not.
- Combine it with other activities such as reading, conversation and projects.
Software can be a useful helper in tailoring practice, yet it only sees part of the picture. The learner, parent or teacher is still the best judge of whether it is helping.