AI enters the classrooms: how to use it to encourage critical thinking in students

Artificial Intelligence has radically transformed our reality in many aspects, but it has had a particularly strong impact on young people and in education in particular.

Oliver Thansan
Oliver Thansan
02 April 2024 Tuesday 11:14
10 Reads
AI enters the classrooms: how to use it to encourage critical thinking in students

Artificial Intelligence has radically transformed our reality in many aspects, but it has had a particularly strong impact on young people and in education in particular. With advances in machine learning and data processing, AI has optimized the personalization of learning, adapting materials and methodologies to the individual needs of many students.

From virtual assistants that provide academic support to recommendation systems that suggest relevant educational content, AI is redefining the way we teach and learn, offering unprecedented opportunities for educational access and innovation.

One of the tools that has been heard the most about has been Chatgpt, especially used by some students and whose use has generated the need to establish standards among the educational community to differentiate the work of students from that of AI.

Regarding the use of AI, the professor at the UPC School of Telecommunications and Aerospace Engineering of Castelldefels, Juan Carlos Aguado, tutor of subjects such as Guided Control, of the Systems Engineering degree (Aerospace Engineering) or Sustainability, of the Engineering degree Aerospace and Telecommunications, explains how he applies AI in class and what his students can use it for, focusing attention on the importance of critical thinking at all times.

What types of AI tools do you use in your classes?

In classes as such I don't use it much at the moment, I mean in front of the students. Related things that I teach them, such as Chatgpt I teach them at the beginning of classes in the Guided Control subject that teaches how to design autopilots for aerospace vehicles. From the degree in systems engineering (Aerospace Engineering). I show it to you so you can see that you can't trust the answers. I give them an example of a typical problem that they could answer and I teach them that they cannot trust the answer.

Chatgpt would be as if you had someone in class who took notes of everything that was said, but didn't understand anything. More or less he will tell you where the shots are going, but you can't trust them. The problem I'm warning you about is that you make mistakes in an unpredictable and quite inexplicable way, at the moment, and they don't know where, but you're going to make a mistake and they won't know when.

I also teach them another tool in the same subject called Wolfram Alpha that takes more time and is not prepared to converse with people, which is what Chatgpt does. It is something more mathematical, which searches for information and its answers are reliable. They are also difficult to interpret, which is why I tell my students that what is given to you in this format you probably won't understand it, but if you change the format you will see that you can understand it.

How do students react to AI in the classroom?

The problem we have is that what we should instill in students at the university is critical thinking. The answers may be wrong and you have to be able to detect the errors. You have to be prepared to be able to, especially nowadays when it is very easy to search for information, interpret it and know when they are trying to place things on you that are false.

I try that, since it is the first day, that this experiment makes them reflect a little and not trust themselves, not think that there are already tools that can give them answers and that can solve problems. At the moment you have to be very aware that what we have is very rudimentary and is not made to be able to solve problems, but rather, to be able to explain it like a person, making mistakes and falsehoods without any problem, because it does not have that concept of what is more or less authentic. But this can change very quickly, because it changes with different versions.

What challenges do you face when using AI?

I warn them that they can use it, that it can be a first approximation, that when they have incomplete information they can at least try to guide them. Nowadays Chatgpt, for example, which is the most common, serves to give you first sketches, approximations. When you don't know how to start doing something it is a good tool. It will give you ideas and some will be quite valid so you can start, others will be wrong and you will not know where the errors are.

Do you think AI affects students' creativity and critical thinking?

Yes, totally. If they use it blindly, if today they believe that it gives them answers that are correct, this completely affects them. It's like having an authority, like an oracle, I don't know why it works or how it works and it will give me answers that I can trust. Today it is quite the opposite. The answers it gives are absolutely unreliable and it is a programming system that has been made to imitate people.

There is a very good definition that I read recently that said that the programs we have now are stochastic parrots. All they do is imitate how people speak and they do it probabilistically. If I have this phrase what is the most likely word that can come next with all the conversations I have in my database.

For what types of educational tasks do your students use AI?

In Guided Control, students have to solve problems and design a type of system that can solve them. Today there are still no systems that do this reliably for you. The ones that are reliable, as I mentioned before, are the ones that do an analysis of the data at the beginning and the data is not 100% reliable at the moment. For that they can use it, but before putting them to tests I first check that we have not reached this situation in which they would be able to solve the problem with AI. Therefore, I know that if they use the tool they will not do it right, they will make mistakes.

What is the goal of using AI?

In this particular subject so far the objective is to indicate to them that there are tools that they can use, that are reliable and that are not going to solve the problem, but that give them input data. And there are tools that at the moment are not at all reliable that can give clues as to where the solution may go, but these tools are going to make mistakes and they will not know where.

Any advice for other teachers who want to incorporate AI into their classrooms?

Above all, they have to be continually updated because, at the moment, it is very unreliable in the answers it gives. Teachers have to be attentive, check what tools they have at their disposal, whether or not they can answer the questions that could be put on an exam and adapt from here. When there are tools that are capable of solving the exams, they will have to change them. This affects all subjects. Right now we already have expert systems that can pass the typical exams to become lawyers by getting a grade of, for example, a 7 or 8 out of 10. It is a matter of paying attention and trying to adapt because everything will change and it will do so very quickly. The use of technology already makes the systems much more powerful in a very short period of time.