Folks waiting for the AI in education bubble to pop have their head in the sand. This isn't coming top down, this is bottom up from the students. Helping students figure out how to not let AI usage harm their learning is now critical. Some things I've been trying below, curious for other ideas:
Comments
https://e-estonia.com/estonia-announces-a-groundbreaking-national-initiative-ai-leap-programme-to-bring-ai-tools-to-all-schools/
How about "Your access to this tool depends on the same corps that cut your access to DRM books and shows. You should
Plus a good dose of practical examples of the tool getting easy answers completely wrong for good measure.
2. I've tried describing these things as harmful to fluency, making you slower in the long term
In the end, it'll be the students who have functionally no experience with prompting LLMs or using AI that end up too slow to compete academically or economically
Educators need to be embracing regular daily use of AI or they're hurting students
4. I've provided examples of where it can lead you highly astray, to help build mental models of where it can mess up
There's no such thing as an unpassable exam: you've merely created a different class of problem to solve.
If they're not using these tools to get a better education they're missing out on a big opportunity in my opinion
You really have work hard to stay deeply curious, so you'll dig deeper.
There is an explosion of new integrated tools which are going to be v powerful too
You can have all the education in the world and never get hired without the degrees
That's why
I think you now have to assume that all out-of-class work might involve some use of AI, then think about how to redesign your course under that reality.
Some ideas in this thread about my Intro to CS class:
https://bsky.app/profile/dansmyers.bsky.social/post/3leieurndss2k