Communications of the ACM
ChatGPT in Computer Science Education
We have all heard it said that ChatGPT and similar applications will dramatically influence all educational systems (see e.g., Nguyen, 2023; Huang, 2023). The question we explore in this blog is how ChatGPT will influence computer science education. We investigated this question in a professional development workshop for Israeli high school computer science teachers that focused on research and entrepreneurship in computer science education.
Interestingly, when we posed the question "How will ChatGPT influence computer science education?" teachers barely addressed the threats it poses (e.g., no need to develop computer programs anymore), but rather highlighted the opportunities that ChatGPT opens for computer science education, as described below.
First, they discussed the basic questions of whether ChatGPT should be integrated into computer science education and whether the computer science high school curriculum should be changed. One of the answers given was that since such tools are presented in the classroom anyway, the question is not whether ChatGPT should be integrated into computer science education, but rather, how it should be integrated into computer science education so as to promote computer science education.
Second, the teachers addressed the various features of ChatGPT. As soon as it was agreed upon that ChatGPT cannot be ignored and should be addressed in computer science classes in some way or another, it was suggested that more attention be given to philosophical aspects of AI and AI ethics, as well as to social issues, such as the future of programming and the computer science profession, and pedagogical questions such as whether or not students should be allowed to submit programs generated by ChatGPT.
Third, ChatCPT was discussed as a tool for enhancing pupils' skills by expanding their knowledge through ChatGPT's answers, fostering their ability to ask questions and to formulate them precisely, and imparting skills to determine the correctness, quality and reliability of ChatGPT's answers, as well as to filter the relevant information received from these answers.
Fourth, the computer science teachers suggested a variety of ChatGPT-based tasks that can be given to computer science students that highlight the message that computer science thinking skills have not become obsolete, despite ChatGPt's ability to answer most computer science questions and tasks explored in school. On the contrary, appropriate computer science tasks can deepen students' understanding of computer science concepts, enhance their computer science thinking skills, and increase the level of abstraction of the discussion. Here are few examples of questions that use ChatGPT, suggested by the computer science teachers:
- Give ChatGPT a programming task and ask the students to explain it.
- Give ChatGPT a programming task and analyze its answer together with the students: Is it correct? How can we check its correctness? Is it readable? Is it efficient? What is its complexity? Does the program work for any and all inputs? How can we test it? Can it be improved?
- Compare ChatGPT's answer to a task with the students' own solutions, find the similarities and explain the differences.
Fifth, the computer science teachers explored ChatGPT as a tool for improving their pedagogy in a variety of ways. For example, in the development of teaching materials (lesson plans, worksheets, and exams), in the implementation of a flipped classroom (in which ChatGPT presents solutions to be explained by the students), and in fostering active learning and self-learning. Indeed, teachers argued that one of the expected products of the integration of ChatGPT into computer science education is a change in teaching methods.
In conclusion, the computer science teachers naturally focused on how ChatGPT can improve computer science education. In fact, their way of thinking reflects Stephen Hawking's message in his book Brief Answers to the Big Questions (2018), in which he discusses the rise of artificial intelligence: "It will either be the best thing or the worst thing, ever to happen to humanity." It seems that the computer science teachers in our workshop saw the glass half-full.
Orit Hazzan is a professor at the Technion's Department of Education in Science and Technology. Her research focuses on computer science, software engineering, and data science education. For additional details, see https://orithazzan.net.technion.ac.il/.
Comments
Udi Raz
In my opinion, the reference to chat GPT in the teaching of computer science is exaggerated. The engine is genius, no doubt. But when I ask myself what abilities I want to impart to the students and what are the ways to achieve them, I find no use for ChatGPT. In the end the chatGPT becomes another type of search engine, one that if it does a good job it actually brings the first good link. I know I'm in the minority...
Cruz Izu
I've recently tested GPT with three tasks related to different teaching topics: writing a simple function, refactoring code and solving a simple number puzzle. He did great with the first, broke the code logic in the second and gave incorrect results on the last.
The puzzle one was interesting - it interpreted the task and proposed a reasonable but incorrect formula, and then it worked out the incorrect result from there. The steps to work out the solution were so bad, students will realise it was wrong, but what about when it "nearly" gets it right? how many teachers and how many students will pick up on the mistakes?
So, GPT will test critical thinking and understanding in subtle ways by being sometimes right and sometimes not so right. Are our students ready to challenge the chat output?
Displaying all 2 comments