AI in Education: Personalized Learning

In a class of thirty, the teacher has to choose one pace. That pace is slow for some and loses their attention, fast for others who fall behind, and only part of the room sits at exactly the right point. This limit has nothing to do with the teacher's ability or the quality of the material; it follows from simple arithmetic, one source of attention against thirty different needs.

Private tutoring has always been the answer to that problem, and it has always been expensive. What has changed in recent years is the ability to reproduce part of that individual attention at scale, not by replacing the teacher but by removing the tasks that consume their time and require no pedagogical judgement.

Three places where machines genuinely help

• Adaptive practice paths: Choosing the next question from previous answers rather than issuing identical exercises to everyone.
• Instant feedback: Automatic marking of structured exercises so learners need not wait until the next session.
• Decline indicators: Signalling to the teacher when someone's activity or scores drift away from their own trend.

Personalization means a different path, not more content

Many education projects confuse the two. Adding more videos and more quizzes increases the learner's load, not their learning. Real personalization means removing what a person already knows and staying on what they do not; the correct outcome is usually less content and more focus, not more. The test is straightforward: if time to mastery has not fallen, personalization has not happened.

What the teacher gains

Marking repetitive exercises, entering grades and chasing absences take a large share of a teacher's time and require none of their expertise. When those tasks are automated, the freed hours go to the work a machine cannot do, which is understanding why a particular student has not grasped a concept. This is why a programme teachers perceive as a threat usually fails, while one that gives them time back is adopted.

Why it matters for Iranian education providers

In the private and online education market, competition for enrolment is fierce, but the main cost sits elsewhere, in attrition. Someone who abandons a course halfway neither buys the next one nor refers anyone, and their acquisition cost is never recovered. Adaptive paths and decline alerts act directly on that point, and unlike advertising their effect compounds.

The first 90 days

• Day 1 to 30: Pick one subject with clear assessment and organize its question bank, tagging each item by difficulty.
• Day 30 to 60: Run the adaptive path with one cohort while another continues with the current method, so comparison is possible.
• Day 60 to 90: Measure completion rate and time to mastery across both cohorts, and extend to the next subject only if they improved.

Common mistakes

• Measuring success by time spent in the platform; more time is not necessarily more learning and is sometimes a sign of confusion.
• Cutting the teacher out of the loop; the system should report to them rather than take their decision.
• A weak question bank; an adaptive path built on poor items merely personalizes the confusion.
• Ignoring the handling of minors' data; guardian consent and a retention policy must be settled before launch.

Three actions for this quarter

• Extract your current course completion rates so there is a baseline for comparison.
• Tag existing questions by concept and difficulty; this is a prerequisite for any adaptive path.
• Hold a weekly session with teachers to review decline alerts so the system connects to a real process.

Frequently asked questions

• Do these tools replace teachers?
  No. They remove repetitive work and direct the teacher's attention to the students who are genuinely stuck.
• Does this apply to vocational and corporate training?
  Yes, and the effect is often visible faster there because skill assessment is more concrete and the outcome more tangible.
• How long should learner data be retained?
  Only as long as the educational purpose requires; write the deletion policy down and communicate it to learners and guardians.

Takeaway

Personalized learning is not a new promise; it is what private tutoring has always offered, now possible at a larger scale. A provider that starts with one subject and one clear metric will learn before its competitors which part of that promise holds up in practice.

Glossary

• Adaptive learning: Adjusting content and pace according to the learner's actual performance.
• Formative assessment: A short check during teaching used to correct course, not to award a final grade.
• Completion rate: The share of learners who continue a course to the end.
• Attrition: Abandoning a course before finishing it.
• Question bank: A collection of items tagged by concept and difficulty level.
• Time to mastery: How long a learner takes to reach a passing level on a concept.