Great post, Jason. Thanks for your urgency! I agree that educators can't just shy away from AI tools and pretend they don't exist. By fall, students will have the world's most advanced language models at their fingertips. But I don't think we've hit an inflection point as yet. For one, hallucinations are a big concern and neither OpenAI nor Google explicitly addressed this issue in their announcements. Teachers are understandably leary of introducing AI that spounts inaccurate information to their students. Teacher discomfort with AI is already high and I don’t expect them to cheerily embrace AI in their classrooms simply because it offers new personalization features for their students. There is still a lot of work to be done in trying to get teachers to look beyond AI simply for automation.
I think there should be a way to balance adapting to AI and pushing against it. I think it's all right if schools decide to take a much different strategy than businesses in adopting (or not adopting) AI, as long as it serves their students.
Yes and we focus a lot on trying to change teachers' beliefs about AI by telling them that they should, when sometimes what is needed is a simple change in practice. I've led many workshops at schools where hesitant or unwilling participants started to believe in the tech only after they were experimenting with it themselves. They came to understand the benefits through their own experiences, rather than because I was overtly attempting to change their belief system. So, when working with a teacher, we need to recognize when to work expressly on their belief system versus when to focus directly on experimentation.
I find it interesting that you seem to portray Khan Academy as outsiders, not educators, and therefor lacking pedagogy. Khan Academy opened in 2008 and as of 2022 had 120 million registered learners in 90 countries. Students have completed over 10 billion practice problems and watched over 1.5 billion instructional videos on the platform. They partner with organizations like NASA, The Museum of Modern Art, The California Academy of Sciences, and MIT. They are non-profit. Of course, Khanmigo isn't perfect, but I'm extremely impressed with what they are trying to do and the number of students they have helped, for free. It is easy to be critical, but I suggest you take a closer look at their mission and what they have accomplished so far.
It's a good point. I don't take anything away from what they have accomplished and what their goals are. I am actually a fan of Khan's.
I do think the pedagogy of their tutor has been very strange, especially in terms of teaching reading. (Admittedly, that could be better with this new version.)
I'm less interested in what they are creating (which I certainly admit is of value) and more in how it has been perceived as a perfect teacher for every student. Part of that is outside of Khan Academy's control. Part of it is because of their marketing.
I am excited about AI tutors. It seems like a good use case. But there seems to be an assumption that they will save education. Things are much more complex than that.
I do think there is a fair point around the fact that these tutors work best for a specific type of student, but maybe more students become that type as they have exposure to better education at a young age. I'm obviously just speculating and being my usual optimistic self. We need to develop curiosity and interest in learning and I think that is all part of my interest in learning games and innovative pedagogies.
I think that is certainly possible! It could go either way. I could definitely see AI tutors like this becoming more effective over time, not just because the tutors get better but because the students own learning processes change because of the prevalence of the tech.
I could also see the opposite happening, as the glow of this tech wears off and AI is no longer engaging by the very nature of being AI.
Great post, Jason. Thanks for your urgency! I agree that educators can't just shy away from AI tools and pretend they don't exist. By fall, students will have the world's most advanced language models at their fingertips. But I don't think we've hit an inflection point as yet. For one, hallucinations are a big concern and neither OpenAI nor Google explicitly addressed this issue in their announcements. Teachers are understandably leary of introducing AI that spounts inaccurate information to their students. Teacher discomfort with AI is already high and I don’t expect them to cheerily embrace AI in their classrooms simply because it offers new personalization features for their students. There is still a lot of work to be done in trying to get teachers to look beyond AI simply for automation.
This is spot on.
I think there should be a way to balance adapting to AI and pushing against it. I think it's all right if schools decide to take a much different strategy than businesses in adopting (or not adopting) AI, as long as it serves their students.
It gets complicated quickly! What's up
Yes and we focus a lot on trying to change teachers' beliefs about AI by telling them that they should, when sometimes what is needed is a simple change in practice. I've led many workshops at schools where hesitant or unwilling participants started to believe in the tech only after they were experimenting with it themselves. They came to understand the benefits through their own experiences, rather than because I was overtly attempting to change their belief system. So, when working with a teacher, we need to recognize when to work expressly on their belief system versus when to focus directly on experimentation.
I find it interesting that you seem to portray Khan Academy as outsiders, not educators, and therefor lacking pedagogy. Khan Academy opened in 2008 and as of 2022 had 120 million registered learners in 90 countries. Students have completed over 10 billion practice problems and watched over 1.5 billion instructional videos on the platform. They partner with organizations like NASA, The Museum of Modern Art, The California Academy of Sciences, and MIT. They are non-profit. Of course, Khanmigo isn't perfect, but I'm extremely impressed with what they are trying to do and the number of students they have helped, for free. It is easy to be critical, but I suggest you take a closer look at their mission and what they have accomplished so far.
Thanks for that, Eric!
It's a good point. I don't take anything away from what they have accomplished and what their goals are. I am actually a fan of Khan's.
I do think the pedagogy of their tutor has been very strange, especially in terms of teaching reading. (Admittedly, that could be better with this new version.)
I'm less interested in what they are creating (which I certainly admit is of value) and more in how it has been perceived as a perfect teacher for every student. Part of that is outside of Khan Academy's control. Part of it is because of their marketing.
I am excited about AI tutors. It seems like a good use case. But there seems to be an assumption that they will save education. Things are much more complex than that.
I do think there is a fair point around the fact that these tutors work best for a specific type of student, but maybe more students become that type as they have exposure to better education at a young age. I'm obviously just speculating and being my usual optimistic self. We need to develop curiosity and interest in learning and I think that is all part of my interest in learning games and innovative pedagogies.
I think that is certainly possible! It could go either way. I could definitely see AI tutors like this becoming more effective over time, not just because the tutors get better but because the students own learning processes change because of the prevalence of the tech.
I could also see the opposite happening, as the glow of this tech wears off and AI is no longer engaging by the very nature of being AI.
I honestly don't know where things will trend.
Spot on!
Thank you!!