Reflection questions for teaching

What needs to be considered when planning teaching in the age of AI?

The following (reflection) questions should serve as an impetus for you to carefully reflect on the design of your courses to ensure that your assessments and examination formats continue to be purposeful and examination-appropriate despite the existence of text-generating AI systems. It is generally important that assessments effectively support students' learning and skills development while taking into account the challenges of modern technology. It is also important that students develop skills in the use of text-generating AI systems.

  • Learning objectives and competencies:
    How do student assessments (partial vs. single act of assessment; formative vs. summative; product vs. process) contribute to the achievement of learning objectives and the development of required competencies? Do they emphasize critical thinking, application of knowledge or deeper understanding?

  • "Resilience" to AI:
    How can exam tasks be designed to be resilient to AI generation? Are there opportunities to take technical or organizational measures?

  • Process-oriented thinking:
    Do the examination tasks promote students' ability to demonstrate the thought process and approach to finding solutions? Could AI-generated answers circumvent this aspect? How could documenting the process help to better demonstrate students' own work?

Origami light bulb
Origami man

  • Authentic performance assessment:
    How can alternative forms of student assessment, such as oral exams, presentations or practical projects, better capture the authenticity of student performance?

  • AI integration in exam and course design:
    Which cognitive tasks do students have to master without AI help? When should students use AI help? Where can AI help lead to better results? (cf. "ChatGPT Advice Academics Can Use Now")

  • "Costs" of using AI:
    Am I (and my students) aware of the "costs" of using AI applications? Even with currently free products, there are follow-up costs, e.g. through data collection from users for the further development of AI systems, high resource consumption through the operation of these systems, ethical "costs", etc. (see HUL "ChatGPT: Case vignette 2")