I am not one to use a tech solution to solve a tech problem (although I admittedly was doing this with the social annotation platform Hypothesis almost a decade ago to promote social annotation and reading accountability—correlation isn’t causation but my best annotators were my highest performing students. Close reading and all that)…but there’s a newish platform out there (Readocracy/Show Your Learning) which I’m likely to deploy this year which again seeks to answer the reading problem.
And here I am deploying a tech solution to a tech problem. Maybe I am one to do that. 🤷♂️
Will take a look. I am not a print snob though I did find that hard copies of chapters for my Great Books class was far preferable to having students use their screens. But I love my Kindle and read a ton online so I'll take a look at the platform. The problem is there is always something new to look at - if I had a nickel for every edtech vendor who pitched me on their latest platform ...
Another threat to reading is that students trust AI to get it right and often don't even read the output. I set them a question, they ask AI, it produces an answer, they copy it and hand it in.
I worked for the last year with secondary school students testing and improving their reading and comprehension. I was shocked how some exceptional readers didn't understand what they were reading. It gave me a different perspective that will strengthen my classroom practice.
AI can be used to help with understanding. You can put a complex sentence or paragraph in and say you need it explained in simpler terms. We need to teach students that there are support options out there and these are available when teachers aren't. We also need to teach them more about AI so they understand the limitations and biases too.
The Brisk story is the honest center of this piece. "I solved the content problem by skipping the reading problem" is a sentence most of us could sign after any given week, and almost nobody says it out loud.
I teach chemistry and physics, so I live on the other side of your science confession. In over 20 years I have watched strong readers hit that same wall daily, stopped cold by domain vocabulary with nothing to hang it on. No strategy substitutes for the base.
One layer to add from my classroom. Reading AI output is not comprehension reading. It is closer to your legal reading. The text arrives fluent, confident, and unsigned, so the student's job is cross examination. The rule I hold my students to is simple. Do not accept the answer until you can verify it and make it your own. A student who can do that reads the model the way you read a brief. A student who cannot is the one the summary reads.
Your focus on reading is right. The stance is the missing half.
My one caveat would be it depends on how sophisticated you are with the prompting - there are AI responses that go on and on and on - I can’t imagine most people - let alone most students - are reading it that carefully. And it’s getting more and more technical depending on how you’re using it. Once you start wading into the coding arena, it becomes another layer of vocabulary to master.
Agreed, and the length problem cuts the same way. A response that goes on and on is not more helpful, it is more to verify, and most readers quietly stop verifying somewhere in paragraph two. Teaching students to demand shorter output is itself a reading skill. Thanks for the piece. It gave me plenty to carry into the fall.
I almost stopped dead in my tracks when I read that you read six hours a day?! I don’t think I’ve read that much since I was working on my undergraduate thesis. These days I’m lucky if I get 20 minutes reading children’s books with my daughter. I’m glad my parents were voracious readers (mom was a media specialist).
Perhaps the honest answer is that AI reading tools can be useful for adults who already have a lifetime of reading behind them, but genuinely risky for students who don’t yet have enough experience to judge the output critically.
Thanks Stephen but your post is too long to read. Could you compress the main points into 250 words please?
Ha!
I can do you one better ...
https://docs.google.com/presentation/d/12iveSa8if2Z_-lcCeg5xZyPXqb3KnClj4CNqNiXuYyQ/edit?usp=sharing
I am not one to use a tech solution to solve a tech problem (although I admittedly was doing this with the social annotation platform Hypothesis almost a decade ago to promote social annotation and reading accountability—correlation isn’t causation but my best annotators were my highest performing students. Close reading and all that)…but there’s a newish platform out there (Readocracy/Show Your Learning) which I’m likely to deploy this year which again seeks to answer the reading problem.
And here I am deploying a tech solution to a tech problem. Maybe I am one to do that. 🤷♂️
Will take a look. I am not a print snob though I did find that hard copies of chapters for my Great Books class was far preferable to having students use their screens. But I love my Kindle and read a ton online so I'll take a look at the platform. The problem is there is always something new to look at - if I had a nickel for every edtech vendor who pitched me on their latest platform ...
Useful for digital texts more so than hard copy. But there are some hard copy solutions too.
Nice post, Stephen.
Another threat to reading is that students trust AI to get it right and often don't even read the output. I set them a question, they ask AI, it produces an answer, they copy it and hand it in.
I worked for the last year with secondary school students testing and improving their reading and comprehension. I was shocked how some exceptional readers didn't understand what they were reading. It gave me a different perspective that will strengthen my classroom practice.
AI can be used to help with understanding. You can put a complex sentence or paragraph in and say you need it explained in simpler terms. We need to teach students that there are support options out there and these are available when teachers aren't. We also need to teach them more about AI so they understand the limitations and biases too.
Yes. This is the point I was trying to get at - used effectively, AI can be quite valuable but it requires being able to read well!
The Brisk story is the honest center of this piece. "I solved the content problem by skipping the reading problem" is a sentence most of us could sign after any given week, and almost nobody says it out loud.
I teach chemistry and physics, so I live on the other side of your science confession. In over 20 years I have watched strong readers hit that same wall daily, stopped cold by domain vocabulary with nothing to hang it on. No strategy substitutes for the base.
One layer to add from my classroom. Reading AI output is not comprehension reading. It is closer to your legal reading. The text arrives fluent, confident, and unsigned, so the student's job is cross examination. The rule I hold my students to is simple. Do not accept the answer until you can verify it and make it your own. A student who can do that reads the model the way you read a brief. A student who cannot is the one the summary reads.
Your focus on reading is right. The stance is the missing half.
My one caveat would be it depends on how sophisticated you are with the prompting - there are AI responses that go on and on and on - I can’t imagine most people - let alone most students - are reading it that carefully. And it’s getting more and more technical depending on how you’re using it. Once you start wading into the coding arena, it becomes another layer of vocabulary to master.
Agreed, and the length problem cuts the same way. A response that goes on and on is not more helpful, it is more to verify, and most readers quietly stop verifying somewhere in paragraph two. Teaching students to demand shorter output is itself a reading skill. Thanks for the piece. It gave me plenty to carry into the fall.
I almost stopped dead in my tracks when I read that you read six hours a day?! I don’t think I’ve read that much since I was working on my undergraduate thesis. These days I’m lucky if I get 20 minutes reading children’s books with my daughter. I’m glad my parents were voracious readers (mom was a media specialist).
We got rid of cable many years back and it was the best thing that could have happened ...
Perhaps the honest answer is that AI reading tools can be useful for adults who already have a lifetime of reading behind them, but genuinely risky for students who don’t yet have enough experience to judge the output critically.
This is very good, Stephen. I'm in the 'Bleak House' slow read and endorse your comment!
And also agree re Maryanne Wolf and reading: https://www.juliangirdham.com/blog/attention-2-maryanne-wolf-and-cognitive-patience
This is great! So much to think about.