With AI able to convincingly generate a desired assignment output, some teachers are shifting to process-oriented assessment, some of which involves a student interacting with a teacher-configured AI tool.
AI is forcing an adaptation, but I'm hoping even more critical steps will be taken in modern education. I write about AI in education stuff for a magazine, and loved reading your thoughts Mr. Dasey. Subscribed for more insights!
As I'm moving to some mode of process-oriented instruction and assessment, your post is very helpful. The points you articulate here, including the metacognitive learning that might occur as a result of trying to cheat, are thoughts that I've been starting to have. Thanks for clarifying! Surely we are and must move in the direction of process and dialogical approaches; and not so much because of possible student fakery, but because it's better pedagogy.
Very interesting. I think that this sort of process assessment will ultimately be the way to go.
If we stick to text, I think the faking can be much better than what you'd get from a GPT. I'd be curious to try writing a Python program that did the whole thing. The necessary meta-cognition could be captured by the programmer having insight to the learning process, and there would be a market for faking tools. I wouldn't create a faking tool that I would ever share, but I can see that it is doable.
Shouldn't all this be done in audio, though, rather than in text? There is no AI good enough to fake the real sound of a student learning by exchanging ideas with an AI. This might change, but we seem to be a long way off at present.
Thanks for the post! I think moving to process oriented assessment is an important step, but it is only the first one. More importantly, we will need to move on to dialogic forms of assessment, particularly assessments that are situated in the classroom dynamics in which they take place. I also think that using a Socratic AI tutor is a misguided approach since it does not have access to the lived experiences of the students. An AI-proof assessment method, by its very definition, cannot be based on AI.
To be clear, I am not necessarily promoting process oriented assessment. Depends on learning goals I think, and of course output quality still matters. Nor was I advocating a Socratic set up. This was merely a quick experiment using choices I imagine many teachers would make.
AI is forcing an adaptation, but I'm hoping even more critical steps will be taken in modern education. I write about AI in education stuff for a magazine, and loved reading your thoughts Mr. Dasey. Subscribed for more insights!
As I'm moving to some mode of process-oriented instruction and assessment, your post is very helpful. The points you articulate here, including the metacognitive learning that might occur as a result of trying to cheat, are thoughts that I've been starting to have. Thanks for clarifying! Surely we are and must move in the direction of process and dialogical approaches; and not so much because of possible student fakery, but because it's better pedagogy.
Very interesting. I think that this sort of process assessment will ultimately be the way to go.
If we stick to text, I think the faking can be much better than what you'd get from a GPT. I'd be curious to try writing a Python program that did the whole thing. The necessary meta-cognition could be captured by the programmer having insight to the learning process, and there would be a market for faking tools. I wouldn't create a faking tool that I would ever share, but I can see that it is doable.
Shouldn't all this be done in audio, though, rather than in text? There is no AI good enough to fake the real sound of a student learning by exchanging ideas with an AI. This might change, but we seem to be a long way off at present.
Beautiful assessment! This is well done!
Thanks for the post! I think moving to process oriented assessment is an important step, but it is only the first one. More importantly, we will need to move on to dialogic forms of assessment, particularly assessments that are situated in the classroom dynamics in which they take place. I also think that using a Socratic AI tutor is a misguided approach since it does not have access to the lived experiences of the students. An AI-proof assessment method, by its very definition, cannot be based on AI.
To be clear, I am not necessarily promoting process oriented assessment. Depends on learning goals I think, and of course output quality still matters. Nor was I advocating a Socratic set up. This was merely a quick experiment using choices I imagine many teachers would make.
Very insightful read, Tim! 🌟
I write about humanizing the future of learning. I’d love your insights on my work! 🌸
https://substack.com/@devikatoprani/note/p-177581013
This strikes me as a very productive intuition to follow.