Active vs. Passive Learning: Which Format is More Effective?
Watching a training video can sometimes feel like learning. The content is clear, the explanations flow, and everything seems logical.
But an important question arises: did the learner actually understand what they just watched?
This is one of the main limitations of passive viewing. As long as learners don't have to answer a question, explain a concept, or apply what they've just learned, they can progress through a course without truly identifying what they have mastered.
Interactive learning changes the game. Learners are no longer just receiving content: they ask questions, provide answers, rephrase, make mistakes, receive feedback, and start again.
In this article, we will see why this difference is far from a mere detail for the learner.
Active vs. Passive Learning: What’s the Difference?
Let’s take two learners taking the exact same course.
The first watches a fifteen-minute video. The instructor explains a concept, gives an example, and moves on to the next chapter.
The second listens to the same explanation. But after a few minutes, the instructor asks them a question. The learner must respond. If they get it wrong, they receive an explanation. If they don't understand a concept, they can interrupt the course and ask for clarification.
In both cases, the information transmitted might be identical.
But the cognitive activity required of the learner is not.
This is precisely why research into active learning is so compelling.
A meta-analysis published in Proceedings of the National Academy of Sciences compared 225 studies on traditional lecturing versus active learning in higher education science courses. Students exposed to active methods achieved better results, while the risk of failure was significantly higher in courses based on traditional lecturing.
This does not mean that listening to an explanation is useless.
It simply demonstrates that the explanation becomes more powerful when accompanied by an activity that forces the learner to process what they have just heard.
Why Watching a Course Doesn't Guarantee Understanding
When we follow a fluid explanation, the human brain tends to have a bias, confusing two things:
“I understand what the teacher is saying”
and
“I would be able to recall and use this information on my own.”
These are definitely not the same skills.
As long as the teacher is laying out their reasoning, the learner may feel they are following perfectly. The difficulty arises when they have to reproduce that reasoning without any help.
This is where interaction becomes essential for retention.
A question like: “Can you explain to me in your own words why this rule applies here?” forces the learner to retrieve information from memory and reconstruct it.
In cognitive science, this mechanism is known as retrieval practice. Work by Jeffrey Karpicke and Janell Blunt shows that the act of having to actively retrieve information leads to longer-lasting memory than study methods based primarily on re-exposing oneself to the content.
In other words, answering a question is not just a way to check if you have learned. It can be part of the learning process itself.
Interactions Can Also Help Maintain Attention
Interaction plays a second role: it keeps the learner engaged in the course.
The problem with a video isn't necessarily its quality. Even an excellent video can be watched with fluctuating attention.
A particularly interesting experiment conducted by Karl Szpunar, Novall Khan, and Daniel Schacter examined this phenomenon during the viewing of online courses.
Researchers regularly interrupted a video lecture with short quizzes.
Among the participants who were quizzed regularly, mind-wandering occurred in about 19% of them, compared to 39% to 41% in conditions without these regular tests. The researchers also observed more behaviors that facilitate retention and learning, such as note-taking among subjects in the experiment.
The lesson is important for digital training: an interaction is not just there to make a course more fun. It can have a cognitive function.
When learners know they might be asked a question, that they have to respond, and that their answers have consequences for the rest of the course, it becomes harder to simply let the course play in the background.
Asking Your Own Questions: What Traditional Video Cannot Do
However, there is an even deeper difference between a quiz added to a video and a true conversational experience.
In a quiz, the designer decides the question in advance.
In a conversation, the learner can ask their own.
And this difference is fundamental.
Two students can watch the exact same explanation and not get stuck in the same place.
One might not understand a word.
Another might understand the definition, but not the example.
A third might want to know how the concept applies in a different situation.
A traditional video cannot anticipate all these misunderstandings.
Conversely, a teacher can answer each of their students' questions until the confusion is cleared up.
It is precisely this pedagogical loop that conversational AI now allows us to reproduce in part within digital training.
AI Avatars Turn Courses into Dialogues
A conversational AI avatar is therefore not just about adding a face to a video.
Its true value appears when it becomes a pedagogical interlocutor, like a virtual teacher.
On Complement, the avatar presents the course, but the learner can also ask it questions. They can ask for rephrasing or additional explanation. The avatar can, in turn, quiz the learner, evaluate their answers, and provide feedback. The experience can thus adapt to the answers and needs identified during the journey.
This transforms the digital learning experience.
For example, imagine a learner discovering the principle of data minimization in a GDPR training.
In a classic video module, they listen to the definition and then continue the course.
In a conversational experience, they can immediately ask:
“Does this mean that a company is never allowed to collect data it might need later?”
The avatar answers.
A few minutes later, the roles can be reversed:
“A company asks for a user's full date of birth only to verify they are over 18. Is this consistent with the principle of minimization? Explain why.”
The learner must now think.
Then the AI can react to their answer.
It is no longer just consumed training; the learner is truly participating.
It is this loop, far more than the presence of an avatar on screen, that constitutes the pedagogical value of the device.
Explaining Also Forces You to Structure Your Understanding
Interaction can go even further when the learner has to explain their reasoning.
Work in cognitive science on the self-explanation effect has shown that pushing learners to produce their own explanations promotes a deeper understanding of the content studied.
This is an important nuance.
A good pedagogical interaction, therefore, does not only ask:
“A, B, or C?”
It can ask:
“Why did you choose this answer?”
or:
“Explain this concept to me in your own words.”
The learner is no longer simply evaluated on their ability to recognize the right option. They must organize what they know clearly enough to express it.
This is, once again, much closer to what happens in a one-on-one exchange with a teacher.
78.8% of Students Prefer the Conversational Experience
This difference is also reflected in learner preferences.
During an experiment conducted with IÉSEG, 78.8% of surveyed students stated they preferred the conversational experience to the non-conversational format. *
The qualitative feedback published by the School echoes this: students particularly appreciated the avatar's interactivity, the possibility to ask for additional explanations, and the more human character of the experience.
What they are looking for is not necessarily more content, but the possibility to act on the content.
The Goal Isn’t to Multiply Interactions
Not every interaction automatically improves learning.
An overly frequent quiz can become tedious, while a question that is too simple engages the learner very little. The goal, therefore, is not to add a maximum number of interactions, but to place them at the right time.
The real question is:
“At what point should the learner stop listening and start thinking?”
An effective experience alternates between explanation, questioning, response, and feedback. It thus approaches a natural situation: a teacher who explains and regularly checks that the student understands.
Initial research on AI avatars is moving in this direction. An experimental study published in 2026 with 48 students observed better knowledge scores with an interactive AI avatar than with autonomous online learning. The authors nevertheless call for further research.
Therefore, it is not the AI that teaches on its own, but the way it is integrated into the pedagogy.
From Content You Watch to Content You Learn With
Video remains an excellent way to explain, but it has one limitation: it doesn't know if the person on the other side has actually understood.
Conversational learning adds this missing loop.
Learners can ask a question when they get stuck, ask for rephrasing, be questioned in turn, and receive immediate feedback.
Digital learning is then no longer just about presenting a course.
It begins to react to the learner.
This is where the value of AI avatars lies: not in replacing the teacher, but in extending some of the interactions that make their guidance so valuable.