Blog

AI in Education: what about the teacher’s role?

For many people, artificial intelligence (AI) is still seen as an unpredictable futuristic technology even though it is massively used by digital tools in our everyday lives: social medias, weather application, online shopping, etc. But as far as we know, teachers seem to consider tools that take advantage of AI for educational purpose as pure innovation and their use is still marginalized at this point. In this article, we will focus on all the benefits that AI tools offer to teachers and the pain points which have to be addressed in order to implement such technology on a large scale.

1 – AI benefits for teachers

Artificial intelligence’s purpose is to imitates human cognitive capacities applied on a large scale, by processing a large amount of data that couldn’t be done by humans in a limited time.
Nowadays, digital learning tools generate a huge quantity of data about students, pedagogical resources and teaching impact. But in front of a class of 30 students or more, the teacher will not have enough time to take note of each students’ learning records to analyze the essential information about his level of comprehension and offer the most adapted pedagogical response. AI applied to Education can help profile learners by analyzing the available data to enhance the teacher knowledge about his students. What is their level of expertise on a given subject? Who is most likely to be disengaged? How to best group students together to make them work collectively? These are many teachers questions that AI can help give answers to, based on learning data.

AI potential benefits applied to education don’t stop here! Once we better know each student, adaptive learning algorithms allow to find the adapted learning content to make students work on their weak points and maximize the teaching impact. These recommendations come from the analysis of the student learning data but also from the other students’ behavior, and those from previous years, that AI is able to identify the best learning content and exercises, and to produce recommendations to the teacher.

Once the student has mastered the expected notions, he might forget them if he doesn’t put them into practice on a daily basis. This well-known mechanism of the brain has been highlighted by Hermann Ebbinghaus, an experimental psychologist, back in 1885: we are likely to forget an information when our brain is not stimulated to retain it. It is then important to reactivate the knowledge in the student’s brain for him to consolidate it over the long term, thanks to practical exercises. Artificial intelligence allows to individualize these reminders according to each student capacity of memorization and his level of mastery on given subjects: the student will be more frequently notified on subjects he has trouble to memorize, and the right contents will be adapted to his cognitive profile.

These examples show us that a teacher can take advantage of AI to better understand his class and use it to assist him in his daily work thanks to automatization and massification of the individualization’s approach. Pushed to the limit, AI offers great perspectives when it comes to inclusion of students suffering from language disorders or from a disability situation. Indeed, AI even allows to personalize the pedagogical content as it takes account of mobility or cognitive problems of each student, to prioritize some learning formats among others.

2 – Limits and obstacles to overcome for a massive adoption

Unlike what its name suggests, artificial intelligence is not “intelligent”. AI only works thanks to rules provided by humans or by the analysis of data from the past, to extract models from it. Bias associated to the use of this technology are then significant. These biases may be due to the under-or over-representation of a population in the studied sample, or either come from human and cultural background.

For instance, if algorithms have been trained with learning data from preferred school areas, nothing would guarantee that the recommendations made would be still reliable applied to a disadvantaged area. The same way, if the analyzed data was collected from students without disabilities, the recommendations may be unreliable applied to a student partially deaf or suffering from a motor disability. To avoid these biases, the algorithmic model training must be achieved with the widest possible population in terms of socio-economic differences, disability or other.

It is also necessary that teachers know about biases’ potential risks and consider themselves as the only masters of their pedagogical choices. Tools based on AI can assist them to achieve time-consuming tasks, but human expertise remains the ultimate gate to pass in order to give students educational content. Nowadays, many teachers mistrust the use of these technologies at school because they fear the loss of their ability to take decisions regarding their pedagogical choices. We know that the video projector has gradually replaced the black board, because it could save up teacher’s time as he no longer needed to write down lessons on it, and rather use this time to prioritize explanations and didactic methods. AI must be seen as a tool that releases teachers from some of their tasks, to let them concentrate on those where they are most valuable.

3 – Solutions to get the most out of AI

It is then essential to set up a relationship based on mutual trust and knowledge between humans and machines. French Public Institution of Education, Sports and Research (IGESR) advises to better train teachers in order to use information and communications technology, and distance learning¹. It is advisable that such training also includes general knowledge on artificial intelligence and how algorithms work to establish a necessary trusty relationship about the use and adoption of AI in education. Ethical thoughts on the subject are underway and must continue to be conducted, in education and in other areas that exploit artificial intelligence and data.

The security of students’ personal data must also be ensured. As the Ministry of Education explains on its website², “Subject to rigorous protection, students’ personal data can be used to improve the individualization of learning paths, implement a more effective assessment of their achievements, provide teachers with new teaching tools and school principals with school life services that better meet families’ expectations. “. Teachers are made aware of the main principles of student data protection and data protection officers (DPOs) ensure, in each academy, that those responsible for processing personal data in the academy, schools or establishments, as well as subcontractors and service providers involved in such processing, comply with the GDPR.

Finally, to exploit AI and new technologies to the fullest, it is necessary to create an ecosystem of tools that communicate with each other. Standards exist today and are widely used. This is the case for example with SCORM and xAPI. But they are still incomplete, and work must be done to facilitate the exchange of information between tools, between classes, between institutions, etc.

To conclude:

The deployment of tools based on artificial intelligence, if accompanied by the necessary training and ethical thinking that goes with it, can only be of great benefit to teachers. A better knowledge about students, individualized recommendations for each one and a work of consolidation over long term can improve the learning impact, especially the one about distance learning whose modalities are completely different compared to face-to-face learning. Today we know that learning with digital tools is achievable but can only work with an adapted pedagogy³. Let’s use artificial intelligence to adapt teaching as much as possible to each student, while leaving teachers in charge of their pedagogy.

Sources :

¹ Pedagogical uses of digital technology in a pandemic situation from March to June 2020, governmental source

² Data protection challenges within Education, governmental source

³ Article “En classe, le numérique ne fait pas de miracles“, from French daily newspaper Le Monde.