Abstract:

The importance of acquiring Computational Thinking (CT) skills by a new generation is widely recognized. On the same, the new generation is growing up in the age of Artificial Intelligence (AI), which, in turn, significantly affects how we live and work and solve problems. A natural question arises: what is the CT on the age of AI? From an educational research point of view the question is: "What is the relationship between artificial intelligence and computational thinking?" Research on computational thinking and artificial intelligence is usually carried out separately even though today there is an understanding that both are important to human beings and that they should be introduced in K-12 education.

In order to create a society where each citizen can possess computational thinking skills and participate in the development of the future of AI, there should be a powerful elementary school curriculum that emphasize constructionist learning of artificial intelligence principles using computational thinking competencies.
The proposed research studies computational thinking skills of elementary school students involved in constructionist learning activities based on artificial intelligence methods.

The curriculum is based on three kinds of activities:
1. Using the reinforced machine learning technique for creating a single neuron that implements a logical gate.
2. Training a certain neural network for recognizing specific concepts. This activity requires the student to test and validate results of the training.
3. Training and coding an artificial intelligence engine. Students will insert textual training data and image-based data. The students will also program the client software using a dedicated programming interface to provide the client interaction with the server. In this activity, the researchers will be able to analyze students understanding of the concepts of Classification, Prediction and Generation.

The proposed research will take place as part of structured learning activities of sixth grade students in a teacher-guided whole-class mode. To facilitate the students and teachers with a relatively fast entry to artificial intelligence, the researchers use a variety of online tools amongst them are code.org, MIT Scratch programming environment and the Watson machine learning engine. This will enable scaffolded learning activities involving programming.

As for the research methodology, qualitative and quantitative evaluations are part of the mixed methods research methodology using a variety of evaluation technique, including tests, surveys, artifact-based interviews, in class observations and project evaluations.

Creating artifacts which are able to learn as a learning activity in elementary school is at its dawn. The proposed research may contribute to understanding of computational thinking skills associated with artificial intelligence technologies. Furthermore, it can contribute to practical implementation of the corresponding curriculum in educational systems.

Keywords:

Computational thinking, Artificial Intelligence, Constructionism, Elementary school.