AI is getting better everyday. I still remember using Google's Bard. to try to put together a data dashboard. Most of the formulas it gave me were long and convoluted. I would often have to correct the AI and question the choices it made.
The models have improved so much. Things that I was doing before, are becoming easier to get through. I haven't seen a reason to use a "specialty" AI model — one that is designed for a specific purpose. For at least the tasks I perform, the standard models work really well.
The designers of Ludia came out with an updated version. That got me thinking ... and so I decided to see how a standard AI model would handle UDL feedback for a lesson plan.
It turns out it's better than I expected.
I used Google Gemini — which has become my preferred AI model recently — since they say they do not use data. That's what the policy says at least.
I am thinking of having teachers I work with to first check their lessons for UDL ideas, then meet with me to discuss what ideas they came up with and how they might fit into their lessons.
The sample lesson plan I used:
https://www.commonsense.org/education/digital-citizenship/lesson/what-is-ai
