The Learning Science of AI in Education with Dr. Jeremy Roschelle and Dr. Pati Ruiz
Presented by:
Dr. Jeremy Roschelle and Dr. Pati Ruiz from Digital Promise join the podcast to discuss their learning sciences research into AI's role in education. They share details about an innovative project using AI to improve student reading literacy and explore frameworks for developing AI literacy and responsible use policies in schools.
- Practitioner Toolkit from Digital Promise, provides resources for collaborative learning that are flexible, adaptable, and rooted in real teaching experience
- Challenge Map, from Digital Promise
- U-GAIN Reading, program from Digital Promise seeking to amplify new knowledge about how to use GenAI to create content that matches each student’s interests and strengths, enables dialogue about the meaning of content, and adapts to a student’s progress and needs
- AI Literacy, framework from Digital Promise to understand, evaluate, and use emerging technology
- SceneCraft, program from EngageAI Institute with AI-powered, narrative-driven learning experiences, engaging students through storytelling, creativity, and critical thinking
- As they face conflicting messages about AI, some advice for educators on how to use it responsibly, opinion blog from Jeremy Roschelle
- Teacher Ready Evaluation Tool, helps standardize the way ed tech decision-makers evaluate edtech products
- Evaluating Tech Solutions, ATLIS is an official partner with ISTE to expand the presence of independent school vendors and technology solutions in the Edtech Index
- If you are interested in engaging in research with Digital Promise, or just have a great research idea, share a message on LinkedIn: Jeremy | Pati
More Digital Promise articles:
- GenAI in Education: When to Use It, When to Skip It, and How to Decide – Digital Promise
- Hearing from Students: How Learners Experience AI in Education – Digital Promise
- Meet the Educators Helping U-GAIN Reading Explore How GenAI Can Improve Literacy – Digital Promise
- Guest Post: 3 Guiding Principles for Responsible AI in EdTech – Digital Promise
Transcript
Narrator:
NAIS, welcome to Talking technology with ATLIS,
Narrator:
the show that plugs you into the important topics and trends for
Narrator:
technology leaders all through a unique Independent School lens.
Narrator:
We'll hear stories from technology directors and other
Narrator:
special guests from the Independent School community,
Narrator:
and provide you with focused learning and deep dive topics.
Narrator:
And now please welcome your host, Christina Lewellen,
Christina Lewellen:
Hello everyone, and welcome back to
Christina Lewellen:
talking technology with ATLIS. I'm Christina Lewellen, the
Christina Lewellen:
President and CEO of the Association of technology
Christina Lewellen:
leaders in independent schools. And
Bill Stites:
I am Bill Stites, the Director of Technology at
Bill Stites:
Montclair Kimberly Academy in Montclair, New Jersey.
Hiram Cuevas:
And I'm Hiram Cuevas, Director of Information
Hiram Cuevas:
Systems and Academic Technology at St Christopher school in
Hiram Cuevas:
Richmond, Virginia.
Christina Lewellen:
Hello, gentlemen. How you guys doing
Christina Lewellen:
this afternoon? So nice to see you again.
Hiram Cuevas:
I'm hanging in there doing all right. So glad,
Hiram Cuevas:
enjoying life.
Christina Lewellen:
I want to ask you guys, how are things
Christina Lewellen:
going with AI before we welcome our guests on today, I was kind
Christina Lewellen:
of curious, you guys are in the school year. You've done all the
Christina Lewellen:
back to school fun stuff. Where are we at with AI right now
Christina Lewellen:
taking a temperature?
Bill Stites:
So interestingly, a few things. One, we're in a
Bill Stites:
strategic planning process, and AI is one of our focal points of
Bill Stites:
that. So it's definitely a hot topic. Late last week, we had a
Bill Stites:
meeting with our AI Task Force, and as I mentioned, just before
Bill Stites:
the pod started, I had a meeting with my head of school where we
Bill Stites:
spent a good majority of the time talking about AI. So it's
Bill Stites:
definitely front and center. The one thing that I was happy to
Bill Stites:
have the conversation with my head of school about is that
Bill Stites:
when we talk about AI, we need to be focused on it, not just
Bill Stites:
for how we are using it in the classroom, but how we are
Bill Stites:
thinking about it in the operational areas within the
Bill Stites:
school as well. So I think that those can often be overlooked. I
Bill Stites:
think that those are areas that often try to operate in
Bill Stites:
isolation from the other policies and procedures that
Bill Stites:
govern the school, and making sure that we are clear on how we
Bill Stites:
are talking about AI as a whole school endeavor and not just a
Bill Stites:
an academic endeavor, is A conversation I'm really looking
Bill Stites:
forward to having this year
Hiram Cuevas:
here in Richmond. Gratefully, I think we've
Hiram Cuevas:
surpassed that tipping point where we weren't just talking
Hiram Cuevas:
about it, and we've actually drawn a line in the sand, and we
Hiram Cuevas:
developed a an AI vision statement for our school, and
Hiram Cuevas:
we've also matched it up with an academic integrity policy as
Hiram Cuevas:
well, much like what Bill said, we're looking at this also from
Hiram Cuevas:
the academic side and the operational side, and seeing how
Hiram Cuevas:
we can incorporate it. Given the announcement that Google made
Hiram Cuevas:
this summer about Gemini and notebook LM being a core
Hiram Cuevas:
service, we have opened it up to our faculty and students in the
Hiram Cuevas:
middle and upper schools. Teachers can use it across all
Hiram Cuevas:
three divisions, but we're very excited that everyone feels
Hiram Cuevas:
confident enough that we needed to provide the experience that
Hiram Cuevas:
the boys need to prepare themselves for the college and
Hiram Cuevas:
jobs of the near future. And in addition to those, we are going
Hiram Cuevas:
to be deploying Flint AI as an additional alternative AI bot
Hiram Cuevas:
for us to use.
Christina Lewellen:
That's really cool. It walks in
Christina Lewellen:
parallel. What you guys are talking about with what I'm
Christina Lewellen:
seeing on the road. I've spent some time with trustees and
Christina Lewellen:
leadership teams at independent schools. I have some conferences
Christina Lewellen:
coming up where I'll be speaking to admission professionals and
Christina Lewellen:
heads of school and different audiences at schools. And so I
Christina Lewellen:
think that kind of, I don't even know if it's bifurcated approach
Christina Lewellen:
of both the AI in the classroom for education, but then AI just
Christina Lewellen:
in the operations and in running a school. You know, I'm seeing
Christina Lewellen:
that and talking about that a lot too. So I love the
Christina Lewellen:
conversations are evolving past chat GPT, and I'm really excited
Christina Lewellen:
to welcome our guests today, because I think that we can kind
Christina Lewellen:
of drill into a little bit of this in terms of literacy, and
Christina Lewellen:
in particular the learning sciences piece of this. So guys,
Christina Lewellen:
I'm welcoming to the podcast today, two important guests from
Christina Lewellen:
digital promise. We are welcoming Dr Jeremy Rochelle,
Christina Lewellen:
who is the Executive Director of Learning Sciences Research at
Christina Lewellen:
Digital promise, and we also have Dr Patti Ruiz, she is the
Christina Lewellen:
director of Learning Technology Research at the Center for
Christina Lewellen:
Learning Sciences Research at Digital promise. Both of you,
Christina Lewellen:
welcome to our podcast. We are so, so grateful that you are
Christina Lewellen:
willing to spend your time with us. I'm sure there's a lot of
Christina Lewellen:
demands on your time, so thank you for being with us.
Jeremy Roschelle:
It's great to be here. Christina, so I would
Jeremy Roschelle:
love to start by
Christina Lewellen:
asking you to help our listeners understand
Christina Lewellen:
I think a lot of folks understand the brand of digital
Christina Lewellen:
promise that's a. Brand that gets used and thrown around a
Christina Lewellen:
lot at our independent schools. But can you explain a little bit
Christina Lewellen:
what it means to be learning sciences research at Digital
Christina Lewellen:
promise? Tell us a little bit about the organization and how
Christina Lewellen:
you guys kind of fit into that place.
Jeremy Roschelle:
Let me just start by saying what learning
Jeremy Roschelle:
sciences is, and actually my back story is I was a computer
Jeremy Roschelle:
science student as an undergrad at MIT in the 1980s and I was
Jeremy Roschelle:
totally fascinated by AI, and there was people doing stuff
Jeremy Roschelle:
with AI and education in the 1980s which seems almost
Jeremy Roschelle:
impossible. It wasn't what we're doing today, but people were
Jeremy Roschelle:
doing stuff, but the direction I took it, I was really interested
Jeremy Roschelle:
in people. And how do people learn things. So what the
Jeremy Roschelle:
learning sciences is, it's not necessarily the Educational
Jeremy Roschelle:
Sciences which might be concerned with, like, what's the
Jeremy Roschelle:
best policy for teacher retention or to prevent drop
Jeremy Roschelle:
outs, or, you know, lots of things are important. The
Jeremy Roschelle:
learning science is really concerned with, how do people
Jeremy Roschelle:
learn, and how do we effectively teach? And so it's really inside
Jeremy Roschelle:
the learning process. How do we make that learning process
Jeremy Roschelle:
better? And so that's what we do in our center. And we have about
Jeremy Roschelle:
35 researchers, and we do work at different points, pre K, 12,
Jeremy Roschelle:
up into university, different subject matters. But our
Jeremy Roschelle:
underlying thing is we want to help every learner to learn more
Jeremy Roschelle:
effectively, learn more deeply as they go. And so we try to
Jeremy Roschelle:
apply research methods to do that.
Christina Lewellen:
That's really cool. So what are some of
Christina Lewellen:
the projects that I mean, I can take a guess what you guys are
Christina Lewellen:
working on, but what are some of the projects that these 35
Christina Lewellen:
researchers are focusing on of late?
Jeremy Roschelle:
One of the big projects we have is a R D
Jeremy Roschelle:
center, research and development center funded by the US
Jeremy Roschelle:
Department of Education. And the project is called you gain
Jeremy Roschelle:
reading. And our goal is to take an existing product that listens
Jeremy Roschelle:
to kids as they read out loud, and what the product already
Jeremy Roschelle:
does is it'll notice if they can't pronounce a word or they
Jeremy Roschelle:
struggle in a particular place, and offer them science of
Jeremy Roschelle:
reading based guidance to overcome that particular
Jeremy Roschelle:
struggle. But what we and lots of educators out there observe
Jeremy Roschelle:
is these products have been designed for the typical reader,
Jeremy Roschelle:
but lots of our readers are speaking with an accent or a
Jeremy Roschelle:
dialect or they use different vocabulary, and we're not yet
Jeremy Roschelle:
adapting to fully support English language learners,
Jeremy Roschelle:
multilingual learners and other people whose English isn't the
Jeremy Roschelle:
typical English these products were designed for. And so you
Jeremy Roschelle:
gain reading is all about that. It's about studying the problems
Jeremy Roschelle:
and the solutions to make sure that when we listen to kids,
Jeremy Roschelle:
when a computer listens to kids read out loud, it can really
Jeremy Roschelle:
help them. Patty, you want to add
Pati Ruiz:
something about that. Yes. As part of that project, we
Pati Ruiz:
are working with educator leaders from across 10 districts
Pati Ruiz:
in seven different states, and they're supporting us to help
Pati Ruiz:
develop listening sessions so that we understand what the
Pati Ruiz:
national conversation is and what educators are really
Pati Ruiz:
interested in getting out of these technologies. And they're
Pati Ruiz:
also sharing their thoughts and helping us develop an educator
Pati Ruiz:
toolkit to be able to leverage digital learning platforms to
Pati Ruiz:
support student literacy. So we do much of our work in
Pati Ruiz:
collaboration, not just with developers of ed tech, but also
Pati Ruiz:
with educators who are on the ground and help us bring that
Pati Ruiz:
practical perspective and what actually happens in various
Pati Ruiz:
learning environments into the work that we do.
Jeremy Roschelle:
So you're probably seeing that like with
Jeremy Roschelle:
nape scores just came out really recently, and reading is a huge
Jeremy Roschelle:
area, and the level of proficiency that students are at
Jeremy Roschelle:
varies a lot by the characteristics of those
Jeremy Roschelle:
learners and a lot of AI that's out there right now is
Jeremy Roschelle:
addressing very generic problems, to my taste, like
Jeremy Roschelle:
better lesson plans, supporting kids while they're doing
Jeremy Roschelle:
homework. But reading is such a huge need, and we need very
Jeremy Roschelle:
focused things from AI in order to improve reading that's just
Jeremy Roschelle:
sort of general, hey, I'm gonna point the kid at a chat bot
Jeremy Roschelle:
isn't going to turn them into a kid who's say, in second grade,
Jeremy Roschelle:
just giving them a chat bot isn't going to make them a
Jeremy Roschelle:
better reader. So we're really focused on the specific
Jeremy Roschelle:
processes of how kids read, how expert science of reading based
Jeremy Roschelle:
teachers would tutor them, and how to get those to happen more.
Christina Lewellen:
It's sort of interesting how this is a
Christina Lewellen:
blended project. So I'll say this as a non educator, but I
Christina Lewellen:
spend a lot of time in this space, right? And to look at
Christina Lewellen:
this as looking at something so fundamental as reading and
Christina Lewellen:
language skills, and then supplementing that with an AI
Christina Lewellen:
solution is a. Very interesting blend of the kind of
Christina Lewellen:
conversations we're having right now around AI right like,
Christina Lewellen:
where's the line of making sure that the fundamentals are in
Christina Lewellen:
place so that we can create these very discerning consumers
Christina Lewellen:
of AI as we move forward? And we'll get into that a little
Christina Lewellen:
bit, but I'm just curious if my observation as an outsider is
Christina Lewellen:
something that lands with you? Yeah,
Jeremy Roschelle:
absolutely. The product we're building our
Jeremy Roschelle:
work on, and our partner in this work is Amira learning. When
Jeremy Roschelle:
we're working with schools, the recommendation is that they get
Jeremy Roschelle:
a half hour per week of practice with Amira, which is this AI
Jeremy Roschelle:
based tutoring program. I hope they're doing a whole lot more
Jeremy Roschelle:
reading instruction in grades one through three, than a half
Jeremy Roschelle:
hour per week. So clearly, this is like a hybrid situation, and
Jeremy Roschelle:
in particular, we're expecting the classroom teacher to provide
Jeremy Roschelle:
that basic instruction, but what's hard for the teacher is
Jeremy Roschelle:
to spend enough time with each kid, closely, listening to them
Jeremy Roschelle:
read and noticing exactly what's hard for them and giving them
Jeremy Roschelle:
specific spot advice about that sound or that combination of
Jeremy Roschelle:
letters or whatever it is that's throwing them off. There's like,
Jeremy Roschelle:
1000s of specific skills that you have to learn between
Jeremy Roschelle:
kindergarten grade three, and it's hard for a teacher to be
Jeremy Roschelle:
providing that much practice. And so it's really a supplement.
Jeremy Roschelle:
It's a teacher's assistant, but it's not taking over.
Bill Stites:
So is the AI bot, listening to the students read
Bill Stites:
that are having difficulty that are not native English speakers,
Bill Stites:
and processing what they're saying and offering suggestions.
Bill Stites:
So it's picking up the accent, it's picking up all these
Bill Stites:
different things in order to make these suggestions. Am I
Bill Stites:
correct? Yeah,
Jeremy Roschelle:
that's right. That's exactly it. So the child
Jeremy Roschelle:
is reading a book that's on the screen, and maybe a short
Jeremy Roschelle:
passage depending on how old they are, just a couple words
Jeremy Roschelle:
depending on how old they are, and they're speaking out loud
Jeremy Roschelle:
into a headset and with a microphone so it hear them well,
Jeremy Roschelle:
and it's noticing like, oh, you said the first word fine, the
Jeremy Roschelle:
next word was fine. Then you got stuck. And it might even
Jeremy Roschelle:
diagnose what you said wrong as you did that. And then it's
Jeremy Roschelle:
offering you an exercise, perhaps that will break that
Jeremy Roschelle:
word down in syllables and say, Hey, you're it's the middle
Jeremy Roschelle:
syllable. It's the vowel combination that you're finding
Jeremy Roschelle:
hard here. You're doing well on the consonants, but let's look
Jeremy Roschelle:
at how to say that. Oh, you sound because you're a little
Jeremy Roschelle:
off on that. So it's that kind of thing that it's doing, and
Jeremy Roschelle:
it'll give them practicing, and it'll repeatedly listen to them,
Jeremy Roschelle:
and they get little feedback on the screen when they get it
Jeremy Roschelle:
right.
Bill Stites:
So based on the script that knows what to
Bill Stites:
expect, and then, therefore, when it doesn't hear what it
Bill Stites:
expects. It's not breaking down what it heard. It's just
Bill Stites:
breaking down the word that it was expecting, how to pronounce
Bill Stites:
that word. It's not doing anything with the word that the
Bill Stites:
kid actually said. I mean, it's not listening to that word per
Bill Stites:
se. It's just listening to the fact that what that student said
Bill Stites:
was not what it was expecting in the way in which it was meant to
Bill Stites:
be said.
Jeremy Roschelle:
That's right, you're really close. Bill, yeah,
Jeremy Roschelle:
that's right, it is listening to the kid what they actually say,
Jeremy Roschelle:
but it's comparing it. It knows the text. And the cool thing is
Jeremy Roschelle:
that this technique goes back to the 1990s a guy named Jack
Jeremy Roschelle:
Mostow, Professor Jack Mostow at Carnegie Mellon University,
Jeremy Roschelle:
realized in the 1990s that even though computer speech
Jeremy Roschelle:
recognition was really primitive, if it knew the text
Jeremy Roschelle:
that the kid was trying to read, it would be good enough. And so
Jeremy Roschelle:
also, we avoid the problem of fake content. The texts come
Jeremy Roschelle:
from science of reading experts. The computer is not generating
Jeremy Roschelle:
fake content for the kid to read. This is authorized
Jeremy Roschelle:
content, so we avoid some ethical problems.
Bill Stites:
But it's not a free text reading. It's not like the
Bill Stites:
kids just coming to it and reading. No, that's right. It's
Bill Stites:
a prepared text. It's not like a reading companion that I'm going
Bill Stites:
to read this out loud. No, okay, that makes perfect sense. I was
Bill Stites:
very curious about how this was actually functioning and how the
Bill Stites:
AI learning model was working to do this. But if it knows what
Bill Stites:
the text is, knows what to expect, it makes perfect sense.
Bill Stites:
Thank you for the clarification. So
Hiram Cuevas:
on the flip side, then, from a comprehension
Hiram Cuevas:
perspective, the student who's now going to listen to the
Hiram Cuevas:
computer make the suggestions if they're having difficulty with
Hiram Cuevas:
the pronunciation and just literacy in general, imagine
Hiram Cuevas:
there's additional support in addition to the AI bot to help
Hiram Cuevas:
them with that piece as well with the auditory component of
Hiram Cuevas:
it,
Jeremy Roschelle:
Hiram, you got to come join our research team.
Jeremy Roschelle:
You're so right on the ball. What we're doing is we're going
Jeremy Roschelle:
out in the field. We're watching Real schools, children in
Jeremy Roschelle:
classrooms that have lots of different accents going on, not
Jeremy Roschelle:
hard to find in America, and we're watching what happens, and
Jeremy Roschelle:
we're looking for opportunities to improve these conversational
Jeremy Roschelle:
interactions and ways that would work with a greater variety of
Jeremy Roschelle:
students. And actually, that's where AI gives us a leg up,
Jeremy Roschelle:
because, I'd say, five years ago. So the style of interaction
Jeremy Roschelle:
in a product like this would be, if a student does something, you
Jeremy Roschelle:
give them a very pre configured bit of feedback, and it was
Jeremy Roschelle:
spoken. Amir would do text to speech and would say the thing,
Jeremy Roschelle:
but it can only say that thing that the programmer had put in
Jeremy Roschelle:
before, and now it can be more conversational, and you can
Jeremy Roschelle:
actually engage the student about what they're reading, and
Jeremy Roschelle:
maybe you need to do that for a little bit to keep them
Hiram Cuevas:
interested. It's that guided learning that you're
Hiram Cuevas:
providing, yeah,
Jeremy Roschelle:
it's more like a person, like a parent or an
Jeremy Roschelle:
older brother might do. We're trying to simulate a much more
Jeremy Roschelle:
natural way of interacting around a text
Hiram Cuevas:
and then provide that differentiated instruction,
Hiram Cuevas:
as the various levels within a classroom exist. We're so
Hiram Cuevas:
excited about this. This is exciting. This is very exciting.
Hiram Cuevas:
It's
Christina Lewellen:
really interesting because we jumped
Christina Lewellen:
right into the you gain program and the work that's happening on
Christina Lewellen:
the R and D side. And my assumption is that there's any
Christina Lewellen:
number of problems you could have solved with the use of AI,
Christina Lewellen:
and you happen to walk down this path. So how did you get there?
Christina Lewellen:
In other words, how did you decide that this was going to be
Christina Lewellen:
where you put the effort, energy focus for right now, and you
Christina Lewellen:
think that this is going to be kind of like a bigger gain or a
Christina Lewellen:
better outcome than if you had started someplace else, because
Christina Lewellen:
obviously AI, there's so many tools, so many solutions. You
Christina Lewellen:
all could have put your big brains on any number of research
Christina Lewellen:
projects. So why this one?
Jeremy Roschelle:
Happy to give you a little of that back story.
Jeremy Roschelle:
We also realize we're a group at 35 but that's not enough to boil
Jeremy Roschelle:
the ocean, so we got to pick our problems carefully, so we look
Jeremy Roschelle:
at things like the National Assessment of Educational
Jeremy Roschelle:
Progress for clues, and so one of the things that was really
Jeremy Roschelle:
clear to us, looking at that data in the newspaper, what gets
Jeremy Roschelle:
reported, is just scores are up, scores are down. And, you know,
Jeremy Roschelle:
we know that story, scores are down, but there's a persistent
Jeremy Roschelle:
gap for students who are English language learners, where they're
Jeremy Roschelle:
gaining proficiency at about half the rate of everybody else
Jeremy Roschelle:
and reading. We also know from lots and lots of research, if
Jeremy Roschelle:
you're not reading well by the end of grade three, that has
Jeremy Roschelle:
major consequences for anything else in the curriculum. So it's
Jeremy Roschelle:
not only a big problem in terms of, hey, we're not getting
Jeremy Roschelle:
enough kids to be proficient, but it's also a big problem. If
Jeremy Roschelle:
we don't fix this, there's downstream problems that are
Jeremy Roschelle:
even harder to fix, so that's like a really central place to
Jeremy Roschelle:
try to make a difference. And then realistically, like you're
Jeremy Roschelle:
out in schools, we're out in schools, kids speak with lots of
Jeremy Roschelle:
different accents in our schools, that's just reality.
Jeremy Roschelle:
Teachers are telling us all the time these kinds of software
Jeremy Roschelle:
aren't hearing kids well. So as researchers, we don't want to
Jeremy Roschelle:
just improve a particular product. We want to do things
Jeremy Roschelle:
that would help the whole field. And we think by doing the kinds
Jeremy Roschelle:
of things we're talking with you about, it'll help us prove it
Jeremy Roschelle:
with a mirror learning but the things we learn about how to
Jeremy Roschelle:
work with kids, accents will be shared as general knowledge that
Jeremy Roschelle:
any product in this space could take advantage
Hiram Cuevas:
of. Jeremy, how many different languages does
Hiram Cuevas:
your application detect in terms of the accents? Because I'm
Hiram Cuevas:
thinking of like Northern Virginia schools often have 35
Hiram Cuevas:
different languages represented in their schools.
Jeremy Roschelle:
Yeah, again, you'd be a great researcher on
Jeremy Roschelle:
our team. We would just turn around say, how many do we
Jeremy Roschelle:
really need to detect? Like, how many meaningful variations are
Jeremy Roschelle:
there? So interestingly, I'll just give you an interesting
Jeremy Roschelle:
example. Southern is an accent that's important to detect. The
Jeremy Roschelle:
experts we're working with on this project come from the
Jeremy Roschelle:
University of Pennsylvania, and they're really talented at this
Jeremy Roschelle:
process, and we're trying to empirically figure out what are
Jeremy Roschelle:
the number and kinds of ways to group students so that we can
Jeremy Roschelle:
more effectively intervene with them. So there may be a whole
Jeremy Roschelle:
Arabic cluster, for example, may not need to do every single
Jeremy Roschelle:
language in Spanish. There might be a couple of different
Jeremy Roschelle:
variants that are really important to pay attention to
Jeremy Roschelle:
absolutely and so on and so forth. But it's an empirical
Jeremy Roschelle:
problem, and we're chipping away at it. Do you handle Dyslexia as
Jeremy Roschelle:
well? To some extent, but there is a parallel center. There were
Jeremy Roschelle:
2r and D centers like this funded, and there's one out of
Jeremy Roschelle:
the University of Buffalo that is more about that kind of
Jeremy Roschelle:
thing. But these are all related. They're just different
Jeremy Roschelle:
ways of speaking. And so what we're trying to do is to go from
Jeremy Roschelle:
a system that does a really cool thing, listen to kids read out
Jeremy Roschelle:
loud and give them feedback, and then, hey, but kids are not all
Jeremy Roschelle:
the same, so how do we deal with learner variability?
Christina Lewellen:
So can you help me understand how the R and
Christina Lewellen:
D piece fits into digital promise?
Hiram Cuevas:
Before I do,
Jeremy Roschelle:
I want to mention something. So we got
Jeremy Roschelle:
really excited about you gain reading and we might move on.
Jeremy Roschelle:
What I want to mention is our ability to do. What we're
Jeremy Roschelle:
talking about depends on work with schools and getting data
Jeremy Roschelle:
sharing agreements so we can get examples of student voices. So
Jeremy Roschelle:
if there are listeners out there who are school superintendents
Jeremy Roschelle:
or heads of literacy, and they get really excited about this,
Jeremy Roschelle:
we're always looking for schools to work with that have
Jeremy Roschelle:
interesting populations with a lot of variants, and that would
Jeremy Roschelle:
be willing to work with us on the data sharing agreements so
Jeremy Roschelle:
we can hear their students
Christina Lewellen:
voices? Yeah. I mean, one of the things
Christina Lewellen:
I would think is that we have a lot of boarding schools within
Christina Lewellen:
our community, and some of those international students come in
Christina Lewellen:
to the states, and I would imagine that there's some
Christina Lewellen:
opportunity there. So we can definitely leave all of this
Christina Lewellen:
call to arms in our show notes and also as part of our
Christina Lewellen:
conversation, because it's possible that there are subsets
Christina Lewellen:
of independent schools that might be a really great fit for
Christina Lewellen:
you and your team.
Jeremy Roschelle:
Yeah, that would be super cool. We'll put
Jeremy Roschelle:
it in the show notes. Okay, so how does research fit into
Jeremy Roschelle:
digital problems? Really well. I'm happy to report it's the
Jeremy Roschelle:
main thing. It's such a pleasure to be at an organization that so
Jeremy Roschelle:
is interested in and respects research, and since digital
Jeremy Roschelle:
promise does a lot of work that is out there, directly
Jeremy Roschelle:
supporting schools like our league of innovative schools is
Jeremy Roschelle:
a great example. Through that work, we get to learn about what
Jeremy Roschelle:
are the problems really facing school leaders, facing teachers,
Jeremy Roschelle:
parents, students. The league, for example, produces their map
Jeremy Roschelle:
of problems from digital promise, we get the sense of
Jeremy Roschelle:
what are the important, pressing problems, and that helps us
Jeremy Roschelle:
figure out what kinds of interesting research questions
Jeremy Roschelle:
we can ask. We also do a lot of translational work. And what I
Jeremy Roschelle:
mean is, you know, researchers speak in jargon, and they don't
Jeremy Roschelle:
present things in ways that educators can pick up and use.
Jeremy Roschelle:
So I'll toot the horn of one of my colleagues, Judy Fusco, who's
Jeremy Roschelle:
just done an amazing thing. There's a huge body of research
Jeremy Roschelle:
on collaborative learning and how to really support kids in
Jeremy Roschelle:
collaborative learning, and that's something on every
Jeremy Roschelle:
schools portrait of her graduate, you're going to find
Jeremy Roschelle:
collaborative learning, but they can't read the research handbook
Jeremy Roschelle:
to figure out what they should know from research 2000 pages
Jeremy Roschelle:
and in horrible jargon. And so Judy worked with a team of
Jeremy Roschelle:
educators to figure out, what are some high value concepts
Jeremy Roschelle:
from that research. How could we express them in ways that
Jeremy Roschelle:
educators can grab hold of them, and also even produce some
Jeremy Roschelle:
videos showing what it looks like in practice to do the
Jeremy Roschelle:
research based thing. And so we do a lot of that kind of work
Jeremy Roschelle:
too, where, where we're doing original research to study
Jeremy Roschelle:
problems that schools care about, and we're doing some
Jeremy Roschelle:
research translation, I guess I would say if we do a third thing
Jeremy Roschelle:
too, because it ties into this topic of AI, and this is where
Jeremy Roschelle:
potty can come in. For sure, is that we had the opportunity to
Jeremy Roschelle:
be a little bit ahead on the topic of AI, because we were
Jeremy Roschelle:
already studying it before it went big. And so we also offer
Jeremy Roschelle:
our expertise and advice and help policymakers. So maybe
Jeremy Roschelle:
that'd be a great place for you to say some more potty bits in
Jeremy Roschelle:
the work we're doing. That's more policy.
Christina Lewellen:
Yeah, I'd love to hear your perspective,
Christina Lewellen:
especially because I think that from the perspective of
Christina Lewellen:
independent schools, we tend to lean in to AI and
Christina Lewellen:
experimentation, as you heard at the top of the show. I'm asking
Christina Lewellen:
the guys like, how's it going at your school this year? And
Christina Lewellen:
they're walking a path. But I would love to just kind of hear
Christina Lewellen:
your perspective and the work that you're doing with schools,
Christina Lewellen:
and kind of where you think the industry, the educational space,
Christina Lewellen:
is with AI.
Pati Ruiz:
Yeah, happy to talk about how we've worked with
Pati Ruiz:
district leaders, especially when generative AI came out, we
Pati Ruiz:
did a lot of work with district leaders to help them be
Pati Ruiz:
responsive and proactive about how they were allowing for use
Pati Ruiz:
in their schools and districts. And it was at a time when there
Pati Ruiz:
was still a lot of uncertainty about how generative AI was
Pati Ruiz:
really going to be implemented in classrooms or not, and so we
Pati Ruiz:
worked with a group of League of innovative school district
Pati Ruiz:
leaders to help develop and identify the key topics that
Pati Ruiz:
they were interested in addressing in their responsible
Pati Ruiz:
use policies. I came into it as a researcher, and I'll just give
Pati Ruiz:
you a little bit of my background, because I did work
Pati Ruiz:
in independent schools, and shout out to joy Lopez at Sacred
Pati Ruiz:
Heart schools in Atherton, who is one of your members, because
Pati Ruiz:
she taught me a lot of what I know about putting technology
Pati Ruiz:
and pedagogy together to Ensure that the technology is
Pati Ruiz:
supporting learners and focusing on their needs and the needs of
Pati Ruiz:
teachers. So we identified five key topic areas within that
Pati Ruiz:
group about priorities that district leaders needed to
Pati Ruiz:
include in their guidance to their teachers, so that they.
Pati Ruiz:
Could support not only the students, but also we hear a lot
Pati Ruiz:
of parents coming directly to the schools and to the teachers
Pati Ruiz:
for guidance. So the whole community really needed more
Pati Ruiz:
support. So we have those topics which are safety and
Pati Ruiz:
prioritizing student, teacher and community data, security and
Pati Ruiz:
safety, the ethics side of it, transparency about the process
Pati Ruiz:
from school leaders and acknowledging that this
Pati Ruiz:
technology is changing rapidly, and there might need to be edits
Pati Ruiz:
to these policies, and that we're all learning together, and
Pati Ruiz:
then a key focus on professional learning and guidance for
Pati Ruiz:
implementation, and not just at the various grade levels, but
Pati Ruiz:
also in different subject areas, as Jeremy was speaking earlier,
Pati Ruiz:
there are specific things that are needed in an English
Pati Ruiz:
language arts classroom, and there are different needs in
Pati Ruiz:
math classroom computer science classrooms, so being responsive
Pati Ruiz:
to those needs, and then regularly evaluating these AI
Pati Ruiz:
systems and tools to better understand whether they're
Pati Ruiz:
benefiting learners causing harm, and what can be done to
Pati Ruiz:
make modifications and ensure that they are working for the
Pati Ruiz:
learners. One of the big pieces that came out of that work is
Pati Ruiz:
the need to focus on professional learning, and
Pati Ruiz:
specifically AI literacy for both teachers and learners. And
Pati Ruiz:
so that's a big part of the work that I've done at Digital
Pati Ruiz:
promise. And as Jeremy said, we've been working in this space
Pati Ruiz:
for a very long time, and we've always taken a human first
Pati Ruiz:
approach and an approach that puts people first. So we think a
Pati Ruiz:
lot about envisioning safe, equitable and meaningful AI
Pati Ruiz:
enabled learning environments through the development of AI
Pati Ruiz:
literacy, which for us, means developing the knowledge and
Pati Ruiz:
skills that enable humans to critically understand, use and
Pati Ruiz:
evaluate AI systems and tools to safely and ethically participate
Pati Ruiz:
in our increasingly digital world, and we have a whole
Pati Ruiz:
framework that I can point you to and Put in the show notes.
Christina Lewellen:
Do you think that schools are landing with
Christina Lewellen:
that framework? In other words, is it resonating with them? Is
Christina Lewellen:
it fill in the need at the moment? Because obviously, I
Christina Lewellen:
can't speak for public schools and districts, but for the
Christina Lewellen:
Independent School world, we are all over the map in terms of
Christina Lewellen:
adoption comfort level, whether we're using it in the classroom
Christina Lewellen:
at all or just using it on the productivity efficiency side.
Christina Lewellen:
And so I'm just curious, do you think that as this all unfolds,
Christina Lewellen:
so too will the framework I would imagine there might be an
Christina Lewellen:
evolution there?
Pati Ruiz:
Yeah, just like you heard me say earlier, we know
Pati Ruiz:
that these things need to be co designed with the communities
Pati Ruiz:
and specifically the end users of these so one of the things
Pati Ruiz:
that Jeremy and I have been exploring is partnering with
Pati Ruiz:
districts and school systems, larger school systems, in order
Pati Ruiz:
to develop a version of this framework that works within
Pati Ruiz:
their context and works within their needs, so that the AI
Pati Ruiz:
literacy practices that they identify are the ones that go
Pati Ruiz:
along with the portrait of a graduate that we mentioned
Pati Ruiz:
earlier as well, because we know that that's where the school or
Pati Ruiz:
district is focused, and We think that these things need to
Pati Ruiz:
be complementary. And of course, while centering the core values
Pati Ruiz:
of this framework, which to us are human judgment and centering
Pati Ruiz:
justice in the use of AI systems and tools, but absolutely, we
Pati Ruiz:
encourage folks to take our framework and adapt it to meet
Pati Ruiz:
their needs.
Jeremy Roschelle:
I would think of the framework sort of as a
Jeremy Roschelle:
way to think carefully about the problem, but not prescriptive.
Jeremy Roschelle:
Terribly. You might come in and your first thing is, I want to
Jeremy Roschelle:
understand AI, and that's what literacy means to me. And
Jeremy Roschelle:
looking at the framework might remind you how important the
Jeremy Roschelle:
evaluate side is, and there's no way we're going to get to safe
Jeremy Roschelle:
AI in schools without really experienced adults all
Jeremy Roschelle:
throughout the school steps and parents who know enough about
Jeremy Roschelle:
how to evaluate these tools. And so we need the evaluation side.
Jeremy Roschelle:
And so it's just a way to make sure you're thinking carefully
Jeremy Roschelle:
as you design your own approach,
Christina Lewellen:
so as people who spend a lot of time thinking
Christina Lewellen:
about this, can I ask you each to tell me what concerns you
Christina Lewellen:
about AI in education and also what excites you?
Pati Ruiz:
What excites me is seeing what students come up
Pati Ruiz:
with when they begin. Experimenting with the use of
Pati Ruiz:
tools. For example, a high school student used AI to design
Pati Ruiz:
a uniform for their sports program, and I think that that's
Pati Ruiz:
a great example of using AI to develop something that then gets
Pati Ruiz:
made. So this uniform ended up getting made because the team
Pati Ruiz:
and the student just developed it because they were interested
Pati Ruiz:
in designing a uniform that they would want to wear, and their
Pati Ruiz:
teammates agreed, and they ended up making this uniform. And so
Pati Ruiz:
how learners are partnering with this technology to go beyond
Pati Ruiz:
their current abilities, to be able to extend their capacity to
Pati Ruiz:
be creative and to leverage it as a tool to support their
Pati Ruiz:
goals, right while they maintain that intrinsic motivation right
Pati Ruiz:
to create something new. So that's always exciting to me to
Pati Ruiz:
see how students end up using these tools. And there are a lot
Pati Ruiz:
of cautions that I have for the use of AI, and specifically
Pati Ruiz:
generative AI, primarily the evaluate portion of our AI
Pati Ruiz:
literacy framework, which includes thinking critically
Pati Ruiz:
about the information and the synthetic outputs that are being
Pati Ruiz:
produced, especially by generative AI, and thinking
Pati Ruiz:
critically about the ethics and the impact of those outputs, but
Pati Ruiz:
also of the use of the tool itself. So helping learners and
Pati Ruiz:
teachers understand when they should be using these tools and
Pati Ruiz:
when they should not be using these tools, or identifying
Pati Ruiz:
between use and being okay not using these tools, I think, is a
Pati Ruiz:
critical piece of what we need to help develop in both learners
Pati Ruiz:
and teachers,
Jeremy Roschelle:
absolutely. And I think that thinking
Jeremy Roschelle:
critically is so important. I mean, one of my worries in this
Jeremy Roschelle:
space is, you know, we go to a lot of conferences, but CNI and
Jeremy Roschelle:
digital promise, and you see the kind of gold rush of companies
Jeremy Roschelle:
rushing to put out products too quickly to get into the market.
Jeremy Roschelle:
Some of them are pretty half baked. And you know, you wonder
Jeremy Roschelle:
how much good they're really going to do. They look cool, but
Jeremy Roschelle:
are they really going to do any good? And they're going to take
Jeremy Roschelle:
time away from other important educational things I worry about
Jeremy Roschelle:
moving too fast and educators not taking it slow enough to get
Jeremy Roschelle:
it right. And on the positive side, the thing is, I've studied
Jeremy Roschelle:
Technology and Learning Sciences my whole professional life, and
Jeremy Roschelle:
we've always had to figure out how to make the technology do
Jeremy Roschelle:
something useful for education, but it was really limited,
Jeremy Roschelle:
especially in terms of conversation, story, discussing
Jeremy Roschelle:
things in the kind of conceptual terms that people talk about
Jeremy Roschelle:
things and AI is really different in opening up new ways
Jeremy Roschelle:
to bring the interaction we can design closer to the research on
Jeremy Roschelle:
the learning and what the learning process should be. You
Jeremy Roschelle:
know, my ideal learning process is one of you guys in a
Jeremy Roschelle:
classroom having a great conversation with a student, but
Jeremy Roschelle:
we can't do that for every student all day long. So if we
Jeremy Roschelle:
could use technology to have some of those great
Jeremy Roschelle:
conversations, that's exciting to me,
Hiram Cuevas:
Jeremy, in terms of the support that this tool is
Hiram Cuevas:
going to have with readers, in terms of, like, a wish list that
Hiram Cuevas:
I could potentially have as a researcher. There are different
Hiram Cuevas:
types of reading. Yeah, there's reading in the humanities,
Hiram Cuevas:
there's reading for science, there's reading in mathematics.
Hiram Cuevas:
Has your organization even gone the adjacent route yet to these
Hiram Cuevas:
other disciplines?
Jeremy Roschelle:
No, but I've thought about it, and it's so
Jeremy Roschelle:
exciting to me, because reading happens across the curriculum.
Jeremy Roschelle:
And it's not true that one's reading challenges are confined
Jeremy Roschelle:
to ELA instruction. And so why shouldn't there be better tools
Jeremy Roschelle:
available when you're a science teacher? I've studied science
Jeremy Roschelle:
education too, and there's a lot of reading difficulties there
Jeremy Roschelle:
because the vocabulary is so tough, and so why don't we have
Jeremy Roschelle:
reading across the curriculum supports, and I think teachers
Jeremy Roschelle:
would welcome those that those would not be interfering with
Jeremy Roschelle:
their role as a science teacher or math teacher. Math, tons of
Jeremy Roschelle:
reading problems come up and prevent kids doing the math
Jeremy Roschelle:
because they can't read the word problem or struggle with
Jeremy Roschelle:
particular words in it. So I think reading across the
Jeremy Roschelle:
curriculum is huge. Likewise, you could do math across the
Jeremy Roschelle:
curriculum and say math shouldn't only be taught in math
Jeremy Roschelle:
class. You're going to come across it in science, in
Jeremy Roschelle:
history.
Hiram Cuevas:
Oh, you're speaking my language, right? But
Hiram Cuevas:
ed
Jeremy Roschelle:
tech products of the past are very curricular
Jeremy Roschelle:
silos, and. This is a more general purpose capability.
Jeremy Roschelle:
Could we use it to do interdisciplinary learning?
Hiram Cuevas:
As a former science teacher, I've often
Hiram Cuevas:
found that the students that struggled most with science
Hiram Cuevas:
often had a language arts deficiency of some sort, and
Hiram Cuevas:
they were really challenged by what I call the vernacular of
Hiram Cuevas:
science. Yeah, you just didn't know the information in terms of
Hiram Cuevas:
communicating as a science student. And likewise, you
Hiram Cuevas:
wouldn't have the same capacity in, say, Spanish or French or
Hiram Cuevas:
Latin, which would also prove to be difficult as well.
Jeremy Roschelle:
You probably experienced this too, and you're
Jeremy Roschelle:
probably, like, one of the top 1% teachers, so you probably
Jeremy Roschelle:
have a great way to handle it, but you'll have science teachers
Jeremy Roschelle:
in a situation, say you're graphing something, and, you
Jeremy Roschelle:
know, students learn something about the equation for that
Jeremy Roschelle:
graph and math class, but you're not the math teacher, so you
Jeremy Roschelle:
don't know exactly how it was taught in math. And wouldn't it
Jeremy Roschelle:
be a great thing to be able to say, hey, you know, I know you
Jeremy Roschelle:
learned about this. Let's make those connections. I'm going to
Jeremy Roschelle:
give a few instructions to your AI assistant, and I want you to
Jeremy Roschelle:
go study those connections between the math you've been
Jeremy Roschelle:
learning in the science I'm talking about here. And like
Jeremy Roschelle:
that kind of connection making is really hard for the average
Jeremy Roschelle:
classroom teacher,
Hiram Cuevas:
and it's so fundamental, I think, to the
Hiram Cuevas:
student understanding well beyond just comprehension, and
Hiram Cuevas:
it's an opportunity for you to actually develop an effective
Hiram Cuevas:
response to the content so that they actually enjoy it.
Jeremy Roschelle:
Absolutely. Yeah, we're all about
Jeremy Roschelle:
stimulating student curiosity about the content they're
Jeremy Roschelle:
learning. That's meaning and understanding. And, you know,
Jeremy Roschelle:
finding ways that AI can give them agency as a learner, to
Jeremy Roschelle:
connect their worlds to what they're learning. Make it
Jeremy Roschelle:
meaningful, make it important. Those are the things we have to
Jeremy Roschelle:
work on. How do we realize those visions?
Christina Lewellen:
So our tech leaders often are in a position
Christina Lewellen:
to support encourage their faculty down this path. How do
Christina Lewellen:
you think tech leaders, you know coming into our core audience of
Christina Lewellen:
CIOs and tech directors? How can they best support teachers right
Christina Lewellen:
now? Part of the context of the question from me is that
Christina Lewellen:
listeners of our pod know that I have a daughter who's a senior
Christina Lewellen:
in college, she's coming through her education program, and she's
Christina Lewellen:
student teaching. She wants to be a teacher, and they're not
Christina Lewellen:
teaching her about AI. Weirdly, program is not really spending a
Christina Lewellen:
lot of time talking about Ed Tech at all, let alone this
Christina Lewellen:
massive tsunami of AI tools that are coming at educators. So what
Christina Lewellen:
that means, I would think, is that the tech leaders in our
Christina Lewellen:
schools that are local, they're not at the district level, so
Christina Lewellen:
they're local, trying to help educators unlock the potential
Christina Lewellen:
while being concerned about the things you're concerned about.
Christina Lewellen:
You know, what can our tech leaders do? Where can they stay
Christina Lewellen:
informed about some of these frameworks, but also just learn
Christina Lewellen:
enough to be able to be great coaches and supporters of the
Christina Lewellen:
faculty who are swamped and busy and all those things, right?
Christina Lewellen:
Well, stay
Pati Ruiz:
tuned. We have an AI literacy in education, PK 12
Pati Ruiz:
brief that we'll be publishing next month, so I will be sure to
Pati Ruiz:
share that with all of you. But as part of that brief, we did
Pati Ruiz:
identify some of these gaps that you're discussing, and the
Pati Ruiz:
recommendations that we give to educators and educational
Pati Ruiz:
leaders are to ensure that there's funding for professional
Pati Ruiz:
learning. We need to be able to give teachers and people
Pati Ruiz:
developing into teachers the time and the resources that they
Pati Ruiz:
need to be able to develop this. This is an ongoing issue with
Pati Ruiz:
professional learning, as we know, but it's also important to
Pati Ruiz:
co create and contextualize the AI literacy policies and these
Pati Ruiz:
frameworks like we discussed with these teachers, so that
Pati Ruiz:
they are able to identify what's important to them and why and be
Pati Ruiz:
better able to integrate these skills, practices and core
Pati Ruiz:
values into their own teaching. It's also important to establish
Pati Ruiz:
educator support systems and help connect educators to one
Pati Ruiz:
another so that they are experiencing more targeted
Pati Ruiz:
professional learning opportunities, right? Depending
Pati Ruiz:
on their grade level subject area. And this can be an ATLIS
Pati Ruiz:
community. I'm sure you have one that you can share in the show
Pati Ruiz:
notes. But also, one final recommendation is we have
Pati Ruiz:
identified a lack of high quality and evidence based
Pati Ruiz:
professional learning and curricular resources when it
Pati Ruiz:
comes to integrating AI, because, as Jeremy said, there's
Pati Ruiz:
a lot that's not fully developed that's out there, and so helping
Pati Ruiz:
educators use. Use evidence based curricula, instructional
Pati Ruiz:
materials, tools and helping to provide guidance on what tools
Pati Ruiz:
they should and should not be experimenting with, I think is
Pati Ruiz:
always a great place to start.
Bill Stites:
One of the things I'm curious about is the idea of
Bill Stites:
targeted professional learning on the website, you've got
Bill Stites:
things listed on there for micro credentialing in terms of, what
Bill Stites:
are some of the things that you offer, what are those micro
Bill Stites:
credentials that you do offer, and how do those, or do those
Bill Stites:
fit into any of the pieces that we're talking about at this
Bill Stites:
time,
Jeremy Roschelle:
the micro credentials, the way that works
Jeremy Roschelle:
is, we're a host for micro credentials, and lots of
Jeremy Roschelle:
different places can develop and post and deliver their micro
Jeremy Roschelle:
credentials through our open source platform. Then within the
Jeremy Roschelle:
platform, we may have specific things. The set of offerings is
Jeremy Roschelle:
changing a lot. So I'm not up to date at this very minute with
Jeremy Roschelle:
exactly what the AI ones are, but I think if educators went
Jeremy Roschelle:
there, they're going to find tons of relevant micro
Jeremy Roschelle:
credential content. And the flip side of micro credentials is
Jeremy Roschelle:
digital promise also does product certifications, and
Jeremy Roschelle:
there are some AI specific product certifications. This is
Jeremy Roschelle:
where companies apply, and they provide a package of materials
Jeremy Roschelle:
saying how we're honoring a principle, like safe AI, and
Jeremy Roschelle:
then we either approve them or go and approve them as actually
Jeremy Roschelle:
honoring that principle. And so that's a good way to you know,
Jeremy Roschelle:
when you're evaluating products, do they have any certifications
Jeremy Roschelle:
from us or somebody else that would give you some confidence
Jeremy Roschelle:
that somebody's done some really deep due diligence on that
Jeremy Roschelle:
product.
Pati Ruiz:
Yeah, and that's a really good point. Jeremy, our
Pati Ruiz:
partners, ISTE, also have the ed tech product evaluation guide
Pati Ruiz:
that can help educators identify well built digital products. And
Pati Ruiz:
then there's also a platform that they host that identifies
Pati Ruiz:
various ed tech products, and connects those to product
Pati Ruiz:
certifications like the ones that Jeremy just mentioned, so
Pati Ruiz:
that teachers or leaders are able to find technology
Pati Ruiz:
solutions that are meeting specific product certifications,
Pati Ruiz:
and that's the Ed Tech index, which I can also share.
Christina Lewellen:
Yeah, ATLIS has also partnered with ISTE on
Christina Lewellen:
that, because we have been aiming to get more independent
Christina Lewellen:
school specific products and solutions on the index. So
Christina Lewellen:
certainly, you know, ed tech tools are fairly agnostic in
Christina Lewellen:
terms of the type of school, but there are some independent
Christina Lewellen:
school specific sis, LMSs and things like that. And we've been
Christina Lewellen:
working as partners with ISTE to try to make sure that, you know,
Christina Lewellen:
we get the representation of independent school solutions as
Christina Lewellen:
well. So we're big fans of that, and we can definitely share that
Christina Lewellen:
information in the notes.
Jeremy Roschelle:
There's two things I would emphasize to
Jeremy Roschelle:
folks who are in the position of guiding adoption or procurement
Jeremy Roschelle:
of AI in schools. Number one is what we all saw during the
Jeremy Roschelle:
personalized learning era, past decade, was schools and
Jeremy Roschelle:
districts take an attitude of, we'll let 1000 flowers bloom,
Jeremy Roschelle:
see what works out there, and then we'll spread the good
Jeremy Roschelle:
stuff. And the problem was the good stuff never got spread, but
Jeremy Roschelle:
all the bad stuff stuck around. And so we are really encouraging
Jeremy Roschelle:
people to be more choosy this round and be more intentional
Jeremy Roschelle:
about what products you're adopting for what reason. In
Jeremy Roschelle:
part, because this is a somewhat riskier technology, so don't let
Jeremy Roschelle:
1000 flowers bloom. Do the hard work to be selective, and the
Jeremy Roschelle:
other thing is to really think not about a one time buy, but
Jeremy Roschelle:
who's your responsible AI partner, who's going to be with
Jeremy Roschelle:
you on this journey? There are companies that I have a lot of
Jeremy Roschelle:
faith in that what could be your responsible AI partner, and
Jeremy Roschelle:
there's others that I just think want to make a sale in this
Jeremy Roschelle:
field, it's going to be so important to only work with
Jeremy Roschelle:
companies that are going to be your partner on this journey.
Christina Lewellen:
That's really great advice. So I'm
Christina Lewellen:
curious, as I play facilitator with our tech directors here on
Christina Lewellen:
the pod, my co hosts and our researchers here from digital
Christina Lewellen:
promise, I have kind of an interesting scenario that I'd
Christina Lewellen:
like to put in front of you guys. So Bill and Hiram, let's
Christina Lewellen:
say that Patty and Jeremy finish up their research that they're
Christina Lewellen:
doing and they're ready to tackle a new grant, a new
Christina Lewellen:
project. What research would you want them to do? You deal with a
Christina Lewellen:
lot of your faculty. You guys are wrestling these issues day
Christina Lewellen:
to day, probably even today, you've been wrestling these
Christina Lewellen:
issues, what do you wish you had more research on at a high
Christina Lewellen:
level, the one
Bill Stites:
area where I would go, and I'm not sure exactly
Bill Stites:
what the research question would be, but I think research at the
Bill Stites:
early learning levels, like when you look at AI, when you look at
Bill Stites:
this type of work, because I think that there is question.
Bill Stites:
About how it is applied, what is appropriate. I taught third
Bill Stites:
grade, and I would always say, if a pen and paper, pencil and
Bill Stites:
paper is better, use that, you know you don't need to use those
Bill Stites:
tools so understanding how, when and why to use AI at those
Bill Stites:
levels. Because, again, I mentioned at the start of this,
Bill Stites:
I came from a meeting with our Head of School, and it was
Bill Stites:
really trying to talk through, how do we articulate that vision
Bill Stites:
statement? How do we articulate the learning goals that we have
Bill Stites:
outlined for AI in the educational sense I mentioned
Bill Stites:
kind of the operational pieces earlier, but from an educational
Bill Stites:
sense, and I think you can talk about some of those things with
Bill Stites:
your middle school and your upper school, your high school
Bill Stites:
age students, but I think it's those early learners where I
Bill Stites:
would really want to see what, how and why we could use it
Bill Stites:
there.
Jeremy Roschelle:
Love it. Bill, we're on it. My co director of
Jeremy Roschelle:
our group, our learning science research group, is heina
Jeremy Roschelle:
Dominguez. She's an expert, and her team is in exactly this
Jeremy Roschelle:
whole early learning area. I think the basic guidance I would
Jeremy Roschelle:
give anyone in this space is people and human relationships.
Jeremy Roschelle:
Caring relationships are so fundamental to young kids. When
Jeremy Roschelle:
we look at the technology, what the research says we often land
Jeremy Roschelle:
on, if you choose a good half hour a week to use technology,
Jeremy Roschelle:
you can see some pretty good gains. So don't go nuts. Find
Jeremy Roschelle:
your best half hour to play with technology, but really double
Jeremy Roschelle:
down on that human dimension of being a great teacher to those
Jeremy Roschelle:
young children.
Hiram Cuevas:
Jeremy, I think what you'd probably hear from
Hiram Cuevas:
the Independent School community is that we are in the business
Hiram Cuevas:
of relationships. Yeah, we've got that pretty much nailed
Hiram Cuevas:
down. My research topic would build off of what Bill just
Hiram Cuevas:
mentioned with the early learners, given that it is so
Hiram Cuevas:
important to build that foundation very early on, is to
Hiram Cuevas:
look at the impact of AI in terms of using manipulatives in
Hiram Cuevas:
science and math education, and how that can be augmented to
Hiram Cuevas:
improve overall literacy with the students in perhaps what I
Hiram Cuevas:
would call the initial Laboratory Sciences.
Jeremy Roschelle:
Yeah, we resonate with that one too. It
Jeremy Roschelle:
relates to work we're doing, absolutely. That's a very
Jeremy Roschelle:
appropriate research topic. A lot is known. I guess my overall
Jeremy Roschelle:
thing about research, since you guys have real research
Jeremy Roschelle:
questions, would be to encourage everyone in the audience to
Jeremy Roschelle:
think about knowledge as really a limiting factor. Right now,
Jeremy Roschelle:
this AI disruption is grabbing us all, making us rethink
Jeremy Roschelle:
things, and it's making us realize what we don't know
Jeremy Roschelle:
research is not just about numbers and finding out how to
Jeremy Roschelle:
count this or measure, that it's about producing useful
Jeremy Roschelle:
knowledge. And so find some friendly researchers who want to
Jeremy Roschelle:
produce knowledge with you and get in the game, because we just
Jeremy Roschelle:
all need to be active growing knowledge in order to use this
Jeremy Roschelle:
technology.
Christina Lewellen:
Well, I love it. That's really cool. Thank
Christina Lewellen:
you for that. Thought. That's a wonderful thing to round out
Christina Lewellen:
this conversation before we let you go. I'm just curious what
Christina Lewellen:
kind of AI tools are open on your desktop right now? What do
Christina Lewellen:
you guys like? What are you using
Pati Ruiz:
open on my desktop now I can say none.
Jeremy Roschelle:
We're actually a little cautious. I'll tell
Jeremy Roschelle:
you. One thing I've been playing with is note takers when they
Jeremy Roschelle:
are going to keep my information safe. But I go to a lot of
Jeremy Roschelle:
meetings, a lot of different people, and some of these note
Jeremy Roschelle:
takers can produce beautiful summaries of the meeting that
Jeremy Roschelle:
are very helpful to me, because I am intrinsically a very poor
Jeremy Roschelle:
note taker, so I'm interested in that. And sometimes I find some
Jeremy Roschelle:
of the research capabilities of AI helpful. I always find it's
Jeremy Roschelle:
nowhere near a real researcher in that area. But if I want to
Jeremy Roschelle:
come up to speed really quickly on the basics of a topic that's
Jeremy Roschelle:
on my periphery, it can be quite helpful to ask for a summary of
Jeremy Roschelle:
recent findings and knowledge in that space I get used to stuff.
Jeremy Roschelle:
Those are my two favorite
Christina Lewellen:
researchers of tomorrow are never going to
Christina Lewellen:
understand the old Dewey Decimal or even Lexus Nexus, right?
Jeremy Roschelle:
Poor John Dewey,
Christina Lewellen:
well, I'd like to thank you both for
Christina Lewellen:
taking the time and spending it with us, because these are big
Christina Lewellen:
picture things that sometimes our tech leaders get into the
Christina Lewellen:
day to day, and they don't take a pause and think about these
Christina Lewellen:
things. We love to see this research that you're doing. I
Christina Lewellen:
can't wait to follow along with your project, and we welcome you
Christina Lewellen:
to come back and talk to us anytime about these types of
Christina Lewellen:
topics, because we'd love to stay tapped into what you guys
Christina Lewellen:
are thinking about.
Jeremy Roschelle:
Thank you for having us. Yeah. Bill Hiram
Jeremy Roschelle:
Christina, it's been a real pleasure.
Narrator:
This has been talking technology with ATLIS, produced
Narrator:
by the Association of technology leaders in independent schools.
Narrator:
For more information about ATLIS and ATLIS membership, please
Narrator:
visit the atlas.org if you enjoyed this. Discussion, please
Narrator:
subscribe, leave a review and share this podcast with your
Narrator:
colleagues in the independent school community. Thank you for
Narrator:
listening. You.
