ChatGPT workshop
The lab helped to co-organize and facilitate a discussion about the role of ChatGPT in our teaching, particularly for those courses that draw upon sociocultural and humanistic approaches to technology in society, such as the courses taught by ETHOS members.
We presented some research that Luis Landa has done into how students are working with the chatbot, mostly in the areas of programming. While many of the examples drawn upon by our University so far to illustrate both the utility and risks to learning and evaluation have been focused on computing education, our discussion reveals some of the risks as well to more qualitative work, specifically written exams that are aimed to discuss, relate, recognize and reflect upon IT systems in real world cases. We shared some examples where we have tested out our own exam questions with the bot to see what it is capable of and then opened up for a discussion on implications for students as well as for organizing our own teaching efforts. There are obvious repercussions to our workload if we keep our exams the same but have to find ways to evaluate and report on suspected uses of the bot that count as cheating. If we want to shift our means of evaluating students, then we also lack resources to, for example, develop alternative exam forms to maintain pedagogic integrity beyond auditing student work for cheating and enforcing repercussions. While some of us may want to creatively enable the use of the bot, especially on courses that pertain to the ethics and impacts of AI systems, many of us feel that there are blind-spots to the bot that could lead to unwanted shifts in the teacher-student relationship.
We discussed how the overall policy of the ITU towards the bot presents a possible entrenchment of OpenAI into our education where we suggest to students that what society needs is people educated in effective use of IT systems, where the bot is one such example.
We also require further resources and strategies to, for example, communicate our expectations regarding the use of AI to our students, understanding the capabilities and biases of the bot, and communicating the risks to students of using the bot as part of their learning. How will this shape learning in the long term? What ecologies of AI might we want in our education beyond this one hyped solution of the moment?