Much is said about the applications of AI in medicine, law, security, marketing, politics and even in the arts, presenting its magic when training an AI algorithm to understand the style of a painter, for example, and generate a painted picture of Chicago according to Munch's style of “The Scream”.

And for education?

AI can contribute significantly to the teaching and learning process. Digital platforms already use AI techniques, such as natural language processing, recommendation systems, cluster analysis, decision trees, prediction and classification, to collect, analyze data and learn student behavior. With that they make personalized recommendations of content and tasks and each student can study according to their pace and difficulties. The consequence of this personalization is to increase student engagement and promote the evolution of their real knowledge.

AI algorithms can predict dropout and identify the main characteristics of students and the institution that impact on this process. This information helps teachers and managers to make decisions before the problem occurs.

AI does not replace the teacher, but assists him in repetitive tasks such as correcting tests and essays, leaving him more free to perform more important tasks. In ENEM, each essay is corrected by two reviewers, when there is a discrepancy between the grades, the review goes to a third reviewer. An evaluating panel makes the fourth correction in the event that the first three have very different scores. The development of an AI algorithm to detect a leak in the newsroom would help to save human and financial resources. If the algorithm detected the leak to the theme, the writing would be sent to an evaluator who would verify this result. Thus, it would not be necessary to hire a second evaluator (Passero et al, 2017).

Understanding the student's profile helps the teacher to act in order to stimulate his learning. Proactive students manage their time, do not procrastinate, are intrinsically motivated and manifest great commitment to tasks (Sinatra, et al 2015).

Motivation and proactivity are usually measured through the application of questionnaires (Bateman & Crant, 1983), (https://selfdeterminationtheory.org/questionnaires/). However, it may happen that the student answers the questions according to the way he would like to be, but not as he really is. Analyzing the data generated from the student's interaction with a digital platform, provides free indicators of desirability.

How to look at a large mass of data and identify, high skills, motivation and proactivity of a class of students?

The figure below shows the number of activities performed by students in the third year of elementary school, as a function of time. These data were extracted from Educacross' digital mathematics learning platform.

The machine learning method known as k-means showed that there are four groups of students, very characteristic: green group whose members perform approximately between 400 and 1200 activities, the yellow group between 200 and 400 activities, the purple group, between 100 and 200 and the blue group between 0 and 100 activities. In the green group, students perform more activities in less time. In the yellow group, the time spent is a little longer, and in the purple and blue groups, there are students who spend a lot of time and perform few activities. Still in the green group there are 3 outliers, students who performed the most activities in the shortest time. These data indicate that the students in the green group are the most motivated and proactive, those in the blue group the least motivated and proactive, and the outliers as well as motivated and proactive are students of high skills.

The great challenge is to find models that can describe the student's profile. There are other characteristics such as degree of metacognition and persistence, which must be extracted from the data. But this is a topic for the next blog!

Finally, another approach to AI in education is as a learning stimulating agent. Students developing codes that use the paradigms of artificial intelligence, can learn concepts from different areas such as calculus, physics, biology.

AI4All is a non-profit organization based in the United States whose goal is to increase diversity and inclusion in Artificial Intelligence education. Teens In AI is an organization whose objective is to give young people from 12 to 18 years old an exposure to the concepts of AI through the creation of a platform for machine learning and data science.

Bibliography

1-https: //ai-4-all.org/

2-https: //www.teensinai.com/

3-https: //selfdeterminationtheory.org/questionnaires/

4-Bateman, T. S., & Crant, J. M. (1993). The proactive component of organizational behavior: A measure and correlates. Journal of organizational behavior, 14 (2), 103-118.

5- Passero, G., Ferreira, R., Haendchen Filho, A., & Dazzi, R. (2017, October). Off-topic essay detection: a systematic review. In Brazilian Symposium on Computers in Education (Brazilian Symposium on Informatics in Education-SBIE) (Vol. 28, No. 1, p. 51).

6- Sinatra, G. M., Heddy, B. C., & Lombardi, D. (2015). The challenges of defining and measuring student engagement in science.