With IDEs like VS Code wanting to autocomplete your entire file based on the first line, how on earth do we write assessment that actually tests the skill of our students, rather than the skills of Claude? This talk will explore what is being done in schools and universities across QLD, and offer suggestions for what you can try in your classroom.
When designing any assessment, we want to ensure it allows students to provide evidence that accurately represents their knowledge, understanding and skills. Ideally, it should also be straightforward to administer and mark!
But programming assessment that may have worked in the past (“create a website for this company”, “program a game with the school mascot”, “query a database to find this information”) is now trivially easy for your favourite LLM to pass without much effort on behalf of the students.
Do you scrap take-home projects and administer all assessment in exam conditions? Do you have take home assessment, but ban GenAI completely, calling it academic misconduct? Or do you go the other way, and openly encourage its use - designing assessment that will be challenging to complete without some sort of coding assistance?
See what is being done and what can be tried next!
Thea is an experienced digital technologies and mathematics teacher from Magandjin/Brisbane. She has taught students from year 4 - university at all different levels of ability, enthusiasm and prior programming experience. Her language of choice is Python (of course) but Thea still insists on going to functional programming meetups so that she can experience what it must be like for her students to sit in a room and understand nothing for an hour. She also co-runs the Python Brisbane meetups - come chat if you want to join!
