Latest / Integrate This! Ed Tech Podcast / Perfect Book Project: When AI Helps Students Write the Books They Want to Read

Transcript

Perfect Book Project: When AI Helps Students Write the Books They Want to Read · Integrate This! Ed Tech Podcast

0:00
Welcome to Integrate This, your go -to EdTech
0:03
podcast where curiosity meets classroom innovation.
0:07
Join your hosts, David Berner and Jeremy Mikla,
0:09
as we explore the rapidly evolving world of educational
0:12
technology. From game -changing tools to insightful
0:15
conversations with today's top EdTech voices,
0:18
we're here to help educators stay informed, inspired,
0:21
and ready to innovate. So tune in, level up,
0:25
and let's integrate this. Thank you for joining
0:41
us for the Integrate This EdTech podcast. As
0:45
you can see, if you're watching this on YouTube,
0:47
we have a very special guest with us here today.
0:50
And if you're not, and you're listening to it
0:53
on either Spotify or Apple podcasts, we thank
0:56
you for joining us. Today we are joined by Liz
0:59
Boeseed. And Liz, we are... So happy to have
1:04
you with us on the program. Liz is an instructional
1:07
technology specialist at Dartmouth Public Schools.
1:12
Her and I were able to connect when she presented
1:15
a poster session at the ISTE Conference in San
1:18
Antonio. Maybe a little over a month ago now.
1:23
It's hard to believe how. fast time flies in
1:26
the summer, but Liz, you know, thank you once
1:29
again. Can you just kind of start off by giving
1:32
us maybe a little bit of your background and
1:34
how did you get into the role in the position
1:36
that you have? Sure. Thank you so much for inviting
1:40
me. I'm excited to do this podcast with you guys
1:43
today. So my name is Liz Vosi and I'm from Massachusetts
1:47
and I work in Dartmouth Public Schools where
1:49
I was a fifth grade teacher for the first 10
1:52
years of my career. And then right before COVID
1:55
in 2019, the position opened up for instructional
1:59
technology specialist. And it was this trial
2:01
by fire because then that spring We went full
2:05
remote and so I learned so much. I went from
2:07
being like the techie teacher to being like an
2:10
actual instructional technology specialist. So
2:13
that hybrid and virtual year taught me so much
2:17
about. personalized learning and reaching and
2:21
teaching our students. And that has really been
2:24
the guide over these past four years and made
2:27
me really fall in love with the position. It
2:29
is the perfect job. It's so much fun because
2:31
I know everyone and I'm able to help everyone
2:34
and I don't evaluate anyone. So it's the perfect
2:38
job for me. It's so much fun and I get to do
2:41
awesome projects like the one that we're going
2:43
to talk about today. Yeah, and that was kind
2:47
of the catalyst a little bit I know Jeremy was
2:49
that is being attended a session with you That
2:53
was about this perfect book project and it it
2:56
seems to combine AI, you know, that's obviously
3:00
a massive topic and then the other side the love
3:03
of reading so those two things Coming together.
3:07
What was the initial spark of inspiration for
3:09
that project? And what were you what were you
3:12
trying to help students do? to be able to engage
3:16
with that. So it's actually really interesting.
3:19
Last summer I was accepted into the Google Innovator
3:22
program, which is a really difficult program
3:27
to get into as an educator. And we spent three
3:30
or four days in Chicago at the Google office
3:32
last summer, really diving into a specific problem
3:35
of practice that we wanted to tackle. And so
3:38
I had chosen as my problem of practice this burden
3:41
of teaching our littlest learners how to live
3:45
in this world of artificial intelligence and
3:49
preparing these elementary teachers to prepare
3:52
those students for that. And so for me, as an
3:54
instructional technology specialist, this is
3:56
a really heavy weight on my shoulders. And so
3:58
through the program last summer, we were able
4:02
to make a plan for different ways that we could
4:06
get some AI literacy for our teachers as well
4:11
as for our students. And I started last school
4:15
year, kind of with my radar up, trying to figure
4:18
out a way to actually integrate it in a meaningful
4:20
way. Because so much about what we do with technology,
4:23
I mean, integrate this. This is what you guys
4:24
are all about. It's not about just tech for tech's
4:27
sake. We have to integrate it and find that meaning,
4:30
that personalization, that differentiation that
4:32
we need to do it. Because last October, I was
4:36
in a literacy meeting, and we've done a full
4:39
swing towards science of reading and really phonics
4:42
-based learning. And our teachers were finding
4:46
that now the books are more accessible for students,
4:49
especially when we talk about decodables, but
4:51
they weren't as interesting. So they were finding
4:54
that the kids weren't engaged in the books because
4:57
when you make a true decodable, you're very limited
5:00
on the vocabulary and the actual words that you
5:02
can use. And so we're sitting there and they're
5:05
like, they just don't like reading these books.
5:07
And that was when I like had this huge light
5:09
bulb moment. I was like, well, that's perfect
5:11
because AI can do that. AI can take a student's
5:15
interest, wrestling, horses, unicorns. And now
5:18
it's got a whole set of words that it can use
5:22
that the student can access plus those phonics
5:25
words. And it's really algorithmic at that point.
5:28
So to put it together to develop a book that
5:30
the student can read and that the student loves
5:32
to read. And so that's when I was like, we have
5:34
something here. And now we need to kind of turn
5:37
it into a project. That's really cool. And I
5:42
love how you use. You did kind of give the nod
5:46
to where not tech just for tech sake, but you
5:49
know something that will be meaningful and and
5:52
really you have to think I think and put at the
5:54
forefront the needs of the learner before everything
5:57
else and then it's once you have that what are
6:00
the tools that you can apply that's going to
6:03
help bring that learning, you know to accelerate
6:06
it to accommodate it to the forefront so that
6:09
it becomes a better experience for or not just
6:12
the students, but the teachers as well. Right?
6:15
Which kind of brings me to this, right? In your
6:18
role, you know, as the instructional technology
6:20
specialist, you're much more, I think, probably
6:24
in front of teachers, more so than your own classroom
6:28
full of kiddos like you probably, like you did
6:31
once have when you started, you know, first out
6:34
of college or whatever it happened to be, but...
6:37
you know and because of that then and I think
6:40
you know, this is sometimes the position we find
6:42
ourselves in too when we don't have our own students
6:46
that we have to find ways to collaborate with
6:49
those classroom teachers. So can you can you
6:53
kind of touch on a little bit about how your
6:56
perfect book project really became a collaborative
6:59
effort between yourself and the teachers that
7:02
you worked with them and and the students as
7:05
well? Yeah, absolutely. So like I said, I really
7:08
have the perfect job. So the way that our model
7:11
is set up is that I see every single class in
7:15
our school every other week with the teacher
7:18
present. So it's student facing, but it's also
7:22
teacher modeling and hands -on like professional
7:25
development. And so the way our model works is
7:28
that there are certain lessons of the year where
7:31
I'm fully in charge. So I cover our digital literacy
7:33
and computer science standards. And then I try
7:36
to work with a teacher during those lessons to
7:38
develop content specific projects that we can
7:41
keep rolling for those other visits that I come
7:44
in. So I'm constantly working with the teachers
7:47
in front of the students. I know all of the students
7:50
and I've been in this role now for six years.
7:52
So I really know all the students and when I
7:56
decided like, okay, we're gonna try to turn this
7:59
into something. I was looking for teachers who
8:03
would be willing to do some risk taking. And
8:07
like I said, I have the best job. So all of the
8:10
teachers in my school are really willing to do
8:13
that. But I found this one third grade teacher
8:15
who I worked very closely with, who was also
8:19
going at the time for her literacy license. So
8:23
she was deep in the world of early literacy.
8:27
And I was deep in this building of an AI foundational
8:32
skills and lessons for students. And so when
8:35
we took our two skills and put them together,
8:38
that's where we were like, OK, let's truly make
8:41
this a collaborative project for the two of us
8:44
as professionals, for the students to be working
8:46
with each other, and then for these two topics.
8:49
So that's where like she was just she was fully
8:52
on board Yeah, that's really great too, I like
8:56
that what you said there about the hands -on
8:59
Approach because I think that rather than you
9:02
know, Liz being like, oh you're in an office
9:05
somewhere doing some things It's like sounds
9:07
like you're right out there with the teachers
9:08
working with them Engaged in the work that they're
9:12
doing in the classroom and that the students
9:13
are doing and really a part of it So that's that's
9:16
really cool What was it, as far as the project
9:19
goes, when you're looking at the AI literacy
9:21
piece of it, how did you determine what skills
9:25
would be important for students to learn about?
9:27
And then what skills did you think would be better
9:30
for the adult teachers to do? Because I know
9:33
there's those two elements to it. From your side
9:37
of things, what did you think would be good for
9:40
you to cover? And then how did you pair that
9:41
with what the teacher would be doing? That was
9:45
a huge challenge. a huge challenge of this project
9:48
was determining really what to and not to teach,
9:52
what they were prepared to learn and not prepared
9:53
to learn because I am all about research -based
9:55
instruction and there was just like last September,
9:59
there was no research on how to teach students
10:02
or even adult learners at that point about these
10:04
skills. But I am lucky that I have had the experience
10:10
of being able to train educators across the country.
10:13
I work part -time for a technology specialist
10:15
as the director of curriculum. So I've been doing
10:19
these training workshops with educators, K -12
10:21
educators across the country. on foundational
10:24
AI skills. And from doing this with such a diverse
10:28
group of adult learners, I think we forget oftentimes
10:30
when we teach teachers, they're also a very diverse
10:33
group of learners with their own needs and goals.
10:36
So I really try to keep that at the forefront
10:38
of my mind when I'm working with educators as
10:41
well as students. And I had found these like
10:44
really foundational pillars that I thought could
10:47
be adapted really to adult learners. and to student
10:51
learners. Because at this point, we're all new
10:53
learners in this topic. It's just that developmental
10:56
ability that you bring to the table. So we really
11:00
started with just this foundation of what is
11:04
machine learning and artificial intelligence.
11:05
And we came up with some really basic simulations
11:08
that we could model for teachers and then have
11:11
the teachers model for the students to teach
11:13
them a little bit about how you train a model.
11:16
and how they're trained and then how they can
11:18
use that training to develop new content. And
11:21
so we did some simulations. And with our first
11:23
graders, that's really where it ended. So the
11:26
first graders were learning about like, OK, how
11:28
are computers developing these stories, these
11:31
words? and we did a simulation called the one
11:34
word story where it's like an old drama exercise
11:38
where you have like if the three of us were together
11:41
we each have our own knowledge base and our information
11:45
that we bring and then we're given a prompt.
11:47
The prompt might be tell a story about a fish
11:50
but then we would go through the three of us
11:52
and each of us could only tell one word at a
11:55
time. while we each have our own knowledge base,
11:58
some of that overlaps, some of that conflicts.
12:00
And if we were to do this exercise over and over
12:03
again, we would continue to come up with a different
12:05
story every time. And then the better the prompt
12:07
is, tell us about a fish who got lost in the
12:10
ocean looking for his dad. Well, then our story
12:12
is going to get more and more predictable. And
12:15
so with our first graders, that's really where
12:16
the lesson ended. And then we talked to them
12:18
about, OK, what would you want this machine to
12:21
develop a story about? And then as we get up
12:24
through second, third, fourth, and fifth grade,
12:26
we start to learn about prompting, feedback,
12:30
refinement, and then really a lot about hallucinations
12:32
and bias as we get up to fourth and fifth grade.
12:36
And it's been great conversations and good foundational
12:40
AI skills for the adults as well as the students.
12:44
Because when I... do these same lessons and activities
12:47
for my adult learners, we also focus on those
12:50
same pillars. We can just kind of go in different
12:53
directions with them. Great. So you're really
12:56
considering that scope and sequence as part of
12:58
it, where you start with, OK, you're playing
13:00
that one word game. And really, that's like an
13:04
exercise in trying to predict what word should
13:06
come next. which is exactly what our large language
13:09
models are actually doing and basing it upon
13:11
their data sets and things like that. But that's
13:13
a great way to introduce to even our littlest
13:16
learners how that model works and could work
13:20
in your own life and how the computers are using
13:22
it as well. And now as you bring this project
13:27
forward, eventually you decided to kind of create
13:32
your own, like a Gemini gem. like your own chat
13:38
bot that you called the Literacy Partner. You
13:41
know, it operates with its own specialized knowledge
13:44
base that includes kind of the components of
13:48
K -12 literacy as part of it. Did you feel like,
13:52
you know, rather than just going to Gemini or
13:55
ChatGPT and just trying to have the kids or the
14:01
adults that were working with them use that that
14:04
large chat bot, why did you eventually settle
14:09
on kind of creating this AI assistant with GeminiGems
14:13
to kind of help you through this project? Do
14:17
you think that was a better way to use your time
14:20
because of that or did you have another reason
14:22
altogether? I'd like to hear what you think.
14:25
No, I mean, that's a great question. And again,
14:27
it's such a complicated answer because if we
14:29
think back to when this project started last
14:31
November, the models and the process have evolved
14:36
so much since then. So I was developing the project,
14:41
I was developing the topics, and I was also developing
14:42
the process and the prompts to be able to get
14:45
a high quality response. And that developing
14:48
the actual prompts, I mean hundreds of hours.
14:53
spent developing these prompts and the process
14:55
to get a high quality book at the end. So I think
14:58
that's an important piece of this is the process
15:00
that the students went through where we started
15:03
by identifying topics within like, what would
15:06
your perfect book genre be? What would the characterization
15:09
be? So they're identifying these specific things,
15:11
but then I needed to craft the prompts to be
15:14
able to take that information. And then like
15:17
you said, apply these literacy foundations. And
15:22
to make it a process where the students could
15:24
not only focus on learning those literary elements,
15:26
but also learn about the AI process. And so what
15:31
they got back from that first interaction, what
15:33
Jim and I provided back to them was five book
15:36
titles, potential book titles, and five synopses.
15:40
And so each of those was also written at the
15:43
student's reading level. So then the student
15:45
went through and picked the one that they wanted.
15:48
the one or two that they liked the best and they
15:51
learned about giving feedbacks and say, what
15:52
do you love about this story? What would you
15:54
want to add? What would like, would you want
15:55
to plot twist? And then, so I had all of these
15:59
prompts based on grade level. I mean, hundreds
16:02
of prompts and they were all in a Google doc
16:04
because gems didn't exist. At least for me, like
16:08
they weren't unlocked for our accounts. I think
16:11
they didn't even go free until like the spring.
16:14
And so I had these Google Docs filled with prompts
16:17
that I would copy and paste over and then copy
16:19
and paste the students' information. And I liked
16:22
being in control of that piece of it. So I did
16:24
it for the pilot class all myself because I wanted
16:26
to make sure there was no personally identifiable
16:28
information at all in any of these. And I also
16:31
wanted to be that content monitor. There's much
16:34
better guardrails now in Gemini than there were
16:36
when this first started. And some of the content
16:38
was concerning based on the students' interests,
16:41
because students have very very broad range of
16:45
interests. So I had these Google Docs just filled
16:49
and filled with prompts. And then when I finished
16:51
the pilot class, I was like, okay, now I would
16:54
really like to be able to teach the teachers
16:56
how to do some of this. And I'm scrolling through,
16:59
like just, I mean. 50 pages of prompts that I've
17:05
now developed based on grade level, based on
17:07
reading level, based on this. And I was like,
17:08
okay, this is not transferable. Like this right
17:11
now is a great me project, but this isn't a project
17:13
that I could hand off to a teacher and say, okay,
17:16
and then you copy this prompt in, and then you
17:18
copy this prompt in. And so that was when they
17:20
had just launched gems for public schools, like
17:25
the EDU accounts. And I was like, I can build
17:28
a gem that holds all of this pre -prompting.
17:32
The one that you saw is really a generalized
17:34
one, but I have specific ones for grade levels.
17:37
For my schools, we use foundations as our phonics
17:40
spectrum. Our decodable gem is fully loaded with
17:44
the actual foundation's phonics spectrum. Once
17:48
you add in the gem, it saves the user from having
17:52
to type in every time, I need an eight sentence
17:55
synopsis, five options. All of those details
17:58
are already there. So then through that workflow,
18:01
it prompts them, OK, what's the student's information?
18:04
And that can even add in like reminders in there.
18:07
Remember, no personally identifiable information,
18:10
like double check for this, make sure they don't
18:12
have their names or the name of their sports
18:14
team. And so then it prompts the teacher to put
18:19
in the information and then prompts them to tell
18:21
them what to do from there. So it's really about
18:24
time saving, but also transferability. But what
18:27
I love about GEMS, I know I'm talking a lot,
18:29
what I love about GEMS is that it's still, all
18:32
that pretext is right there. So it's transparent.
18:35
When we use a lot of these online story generators,
18:38
you don't know what the programmers have done
18:40
for that pre -prompting. So when you put in a
18:42
genre and it throws you back a story, you don't
18:44
know what that other, what other instructions
18:46
it's been given. But with the GEMS, it's totally
18:49
transparent. It's all there if you want to scroll
18:51
through. Right, and I love that part of it and
18:55
I geek out on that a little bit because one of
18:58
my favorite sessions that I will present at different
19:01
conferences or different workshop days is on
19:03
creating your own chatbots. And here basically
19:06
this is what you've done. You've created your
19:09
own chatbots that operate in a way that you want
19:11
to. And I constantly am reminding people and
19:14
it's... If you want it to be good and you want
19:17
it to do exactly what you want it to do, you
19:20
have to keep going back to it and keep iterating
19:22
on it until it is consistently pumping out exactly
19:28
what it is that you want. And it's obvious to
19:31
me that you've definitely spent that time building
19:34
that data set of your own, that knowledge base
19:37
that is then transferable through that particular
19:41
chatbot, whether it's gems or if you're using
19:43
a different one, maybe a notebook LAM one. or
19:46
even chat GPT's ones as well, but really, really
19:50
cool. And I would say on behalf of other teachers,
19:53
thank you for putting the time in to do that.
19:56
Yeah, and it's a great process too to just hear
19:58
you talk about because I think that that's one
20:00
of the things that often gets missed is that
20:04
when you're doing some of these things and working
20:06
with these chatbots or creating these chatbots,
20:09
there's a lot of struggle that goes into it.
20:11
Like it's not just like... Very easy I know you
20:14
know the easy button gets referenced a lot with
20:17
AI But there's so much iteration and things that
20:19
happen in the process of this and I think it's
20:22
also great because you had mentioned that Your
20:25
essentially you're going through that that process
20:28
of struggle so that you can then provide your
20:30
teachers and the opportunity to create some things
20:35
with your chat bot that lift some of the struggle
20:38
off of them because of their time constraints
20:40
and everything and be in front of students all
20:42
day. So that is two things that really stood
20:45
out to me there from what you were detailing.
20:49
As far as the workflow for the project with the
20:53
students, how does the workflow of the project
20:57
contribute to students understanding of AI capabilities
21:00
and then their ability to inspire and develop
21:03
the perfect book? Yeah, so this is another challenge
21:09
of this is that all of my students are under
21:12
13. So there could be no direct AI use by our
21:17
students. And so this was really about teaching
21:21
concepts to students and then having them watch
21:25
and see what those products could be and what
21:30
the process looks like without having firsthand
21:32
experience of it. and we do a lot of digital
21:37
citizenship lessons in our schools and we compare
21:40
going online to being like in a city on a field
21:43
trip. And so I often tell the students like when
21:45
I'm monitoring your accounts or I'm doing these
21:48
things it's I'm chaperoning you on a field trip.
21:50
When we go to a city there are places that it's
21:52
safe for you to go by yourself that are fun and
21:54
educational. There are places that are safe for
21:56
you to go with a grown -up and there are places
21:58
that are just not safe or appropriate for kids
22:00
but we still need to learn about those places.
22:02
So we know when we're ready to go to them what
22:04
what to do there. So this was one of those situations
22:07
where they're not actually going to use it, but
22:09
it's important. We talked to them. It's important
22:11
that they learn about it because it's part of
22:13
their life and a lot of them are experiencing
22:16
it at home when they're outside of the school
22:20
and outside. Even though those age restrictions
22:23
still exist, we know a lot of our students are
22:25
still exposed to it. So coming up with this process
22:29
and being able to teach them how to do it. at
22:33
those various grade levels. So each one builds.
22:35
If you do get a chance to visit the website,
22:38
which I'll showcase in a little bit, it walks
22:40
through the process for each of those grades.
22:43
And so we did it all in Google Classroom. We
22:45
used tabs, which has been life -changing in Google
22:48
Docs since they added tabs. And they started
22:51
off with genre and characterization, and that's
22:55
when they got back those synopsis. Then we learned
22:57
about feedback. And so once the students learned
23:00
how to give solid feedback, what do you like?
23:02
What do you want to change? What do you want
23:04
to add? The teacher at that point was teaching
23:08
about plot elements and story arcs. So that was
23:11
the focus of their feedback, was about building
23:13
a strong, engaging story arc. Then we would put
23:17
that into Gemini with the prompt, so the gem.
23:22
And the student got back then a 10 chapter outline.
23:25
And so the outline was created again to be at
23:28
their reading level, but also to have three events
23:32
in each chapter. And so the student at that point
23:35
was going in and we learned about editing and
23:38
refinement and that you really need to take control
23:41
of this content when we talk about them being
23:43
the developer. And so their job was to add in
23:45
a fourth item to each chapter. So that was either
23:49
an additional event in the chapter, additional
23:51
detail, maybe it was a specific piece of dialogue
23:53
that they wanted in there. And so that's why
23:57
this is really, it's a reading project. It's
23:59
not so much a writing project because the students
24:00
are doing a lot of reading through this whole
24:03
process, all at their reading level and all based
24:06
on a topic that is engaging for them. And then
24:09
it was only at that point did we actually put
24:11
it in to generate the full text because we wanted
24:13
the student to be in the driver's seat all the
24:16
way along. We didn't want to just go in and say,
24:18
write a 10 chapter book about unicorns. We really
24:22
wanted it to be in a way that would be interesting
24:25
and engaging for the student and for them to
24:28
like. Realized and that happened a lot as they
24:30
were collaborating. They're like wait, you have
24:32
a secret door I want a secret door. Wait, you
24:35
have a good guy who turns out to be a bad guy
24:36
I want that and so they like they started adding
24:40
in these story elements and as they realized
24:41
like wait I can add I can fly I can make it so
24:44
I can fly. Yes, I want to fly and so Like as
24:49
the process went on and they were able to realize
24:51
all of these things that they could do It is
24:54
where we end up generating these these incredibly
24:56
unique Stories. Yeah, and I think that's great
25:01
and I think one of the things that you you know
25:04
you're bringing in this human centered element
25:07
to AI at the same time like how you talk about
25:10
where AI gives you three ideas and you've got
25:12
to come up with your fourth and and so it doesn't
25:15
just relinquish all controls over to AI that
25:19
the students do feel like. you know, they've
25:21
got a stake in what the outcome of it is going
25:24
to be, right, that it is still some of their
25:26
own creativity and originality coming into play.
25:30
You just mentioned here, too, that, you know,
25:32
that it's not really a writing project, it's
25:35
a reading project. And I know from conversation
25:38
with you that when students are given credit,
25:43
right, when their book is published, it's not
25:46
as an author. right? I think you described it's
25:50
inspired and developed by. I think those, I'm
25:53
not trying to put words in your mouth, but I
25:55
think that's what you had said. Can you explain
25:58
the distinction between being inspired and developed
26:02
by versus being authored, and how might that
26:05
help students still feel like they're immersed
26:08
into the story, the characters, the plot, and
26:11
the genre rather than just kind of being bogged
26:14
down by the the creation of the text and the
26:17
manuscript and that stuff. Absolutely. That was
26:20
one of the most powerful parts of this whole
26:22
project is like when you look at the library
26:24
as a whole and when you look at these books that
26:26
these students have created, none of these books
26:29
would have been written by an adult for a kid.
26:32
And also none of these kids would have been able
26:36
to have the literary skills to develop these
26:40
very kid ideas into a full book. So these books
26:45
are so unique and personalized to what the student
26:49
wants to read. And it's fostered so many amazing
26:53
conversations with the students, really about
26:56
AI and about keeping that human component. So
27:01
they were working on developing the perfect book
27:03
that they would want to read. Like we say, if
27:05
you walked in the library or the Barnes and Noble
27:07
and there's a glowing book on the shelf calling
27:09
your name, what would that be? And so we had
27:12
a lot of conversations with students as the books
27:14
were being developed and they were like, wait
27:16
a second, like, I don't, like, that's not what
27:18
I would want to, and I would tell them, I would
27:19
be like, that's your book. You go write that
27:21
book and be the author of that book. This is
27:23
just a book that we're writing that you would,
27:25
that you would want to read. But like their stories
27:28
that they wanted to tell, I told them to hold
27:30
onto those and to develop and write them as the
27:33
author of those books. Whereas this was very
27:36
intentionally a book being developed by them
27:39
and inspired by them. for them to read. It was
27:42
really for their enjoyment. Like what is your
27:44
perfect book right now? And I'm so excited for
27:47
next year when I have almost all of these students
27:50
back again and to see what their perfect book
27:53
looks like a whole year later because you develop
27:55
so much and you change so much at that age. I
27:59
think that's going to be a really interesting
28:01
thing for them to have and to double the size
28:03
of our digital library, I think will be amazing.
28:06
They love reading each other's books, even though
28:08
they're not the author of them. That was the
28:11
funny thing when I brought the first pilot books
28:13
into fifth grade. So the pilot class was third
28:15
grade. I brought them into fifth grade and the
28:17
kids are reading them and they're like, these
28:18
are good. Like, I want to read, I want to read
28:22
this. I told them to stop and they were like,
28:23
wait, no. And I was like, well, don't you want
28:25
to make your own? And they're like, oh yeah.
28:28
So that conversation about author versus developed
28:32
by is something that, I mean, these kids are
28:34
gonna be grappling with their whole lives. Yeah,
28:39
and one of the things too is like the promotion
28:43
of student agency through this process and having
28:47
them be kind of the directors of this project,
28:50
right? So in what ways did giving them ownership
28:54
over making the perfect book What kind of ownership
28:59
did that give them as the co -author, I guess
29:02
you could say, and igniting their passion for
29:05
reading and learning? The excitement around this
29:09
project was so palpable in the building. We actually
29:14
had to pause mid -project for state testing,
29:19
because I do all the training and work for our...
29:22
digital state testing and the excitement that
29:26
came when we got back to the book and I walked
29:28
in and they were like, are we testing? And I
29:29
was like, no, we're back to the perfect book.
29:32
they could not wait to see what that next generation
29:35
was what that message it's like it's like when
29:37
you when you when we first got email and you're
29:39
like oh i got an email like someone sent me a
29:42
message like i want it's going way back but i
29:45
remember that feeling of like someone wrote that
29:48
for me and that's that that feeling that these
29:50
kids had like someone took my ideas and turned
29:53
it into something and you can see it grow and
29:57
develop and just their excitement around watching
30:01
their ideas turn into something and seeing the
30:03
potential of what it could be was so exciting.
30:06
And they were also so excited for each other
30:08
and to read what each other like what the other
30:12
students books were and to find students with
30:15
similar interests was also a great one because
30:18
we had a lot of crossover. And they so that now
30:21
gives that student like six books that are on
30:24
a topic that they're interested in that they
30:26
could read. I feel like we've talked a lot about
30:30
the project here, but we haven't really looked
30:32
at any examples of them. And I'm wondering if
30:36
maybe now is a good time that would you like
30:38
to share some of those unique or diverse examples
30:41
of some of the projects that were created? Absolutely.
30:46
Can we pull up that site? Perfect. I'm going
30:49
to click over. So this is the home page for our
30:53
perfect book project. There's a little introduction
30:56
on the front page, and this is for our students
30:59
right now, so for them to be able to use this
31:01
as a tool for summer reading. Sorry to interject
31:04
here really quick. Are you OK if we share a link
31:06
to this so that if people wanted to go view it
31:09
themselves, they would be able to do that, and
31:12
we'll embed that into the show notes and descriptions.
31:15
Great. So this is our digital library. Right
31:19
now it has some instructions for the students
31:21
to use it for summer reading. And then these
31:24
are the links to the actual grade library. So
31:27
right now, the way it works is each of these
31:30
books is really developed for the individual
31:32
students' reading levels. So just by going to
31:34
third grade books doesn't mean they're all at
31:37
the same reading level. And right now they're
31:38
not organized by level, but I'm hoping once we
31:40
double this from 300 to 600 books next year,
31:44
we can get more granular with the filters and
31:46
the way the way that we organize them. And so
31:50
you'll see these are some examples. These are
31:53
the pilot classes books and you can see they
31:55
are inspired and developed by Ben, text generated
31:58
by Gemini. So this is a student who's so interested
32:02
in Batman, loves everything about Batman, and
32:05
was looking for a very unique book about a face
32:09
-off with Two -Face. And he wanted there to be
32:12
a problem where Two -Face would flip a coin,
32:15
and then depending on how the coin would land,
32:17
either something bad or something worse would
32:19
happen. And these students were so excited as
32:23
they developed these books. I mean, Super Sandwich
32:28
is another great one. So the third grade group
32:32
was able to develop their books. And then at
32:34
the end, we talked about a little bit about image
32:36
prompting. And so they were able to use image
32:40
prompting to try to develop the cover image for
32:43
their book. That started in third grade, fourth
32:45
grade and fifth grade were able to do that. And
32:47
I'm hoping that now that image prompting is.
32:50
becoming or image generation is becoming so much
32:53
easier and that we can now streamline and maintain
32:56
the character that next year we'll even be able
32:59
to do illustrations or images for each chapter
33:02
of the book. I have to tell you my favorite book,
33:06
this is a really unique one, was in fifth grade
33:11
and it's actually by a student who loves heavy
33:15
metal music. I like that fifth grader. All right.
33:20
Yes. Yes. And I mean, there are no fifth grade
33:24
appropriate. I mean, there's very limited books
33:27
in general about heavy metal, let alone ones
33:30
that are catered for a fifth grader. So the original
33:34
title of the book was Highway to Hell and Back,
33:37
which is a little dark, but also a really cool
33:40
name for a book. So it's H2H and back is where
33:43
we landed on it. And it is this epic rock battle
33:47
that includes all of the students' favorite bands.
33:50
And the book starts with a legend. I don't know.
33:53
If I open it, will you also be able to see it?
33:55
Let's see. If you don't, you maybe have to switch
33:59
tabs. I think. Did it open? Or no? Yep. It did.
34:05
OK. So it actually, you see each book begins
34:07
with a summary for a student and then some vocab
34:10
words that might be outside of their reading
34:12
level. that they'll find throughout. And when
34:15
the students went through the editing process,
34:16
part of their hallucination hunting was actually
34:18
looking to make sure that all of those vocabulary
34:20
words were actually found in the text. And we
34:23
did find some vocab words that were listed that
34:25
weren't in the text. And so this one starts with
34:29
a prologue of a legend of this. In order to become
34:34
a true rock god, you had to face off against
34:38
the great darkness. And so you can see It starts
34:42
with this legend and then it moves into the chapters
34:46
where all of a sudden there's a concert, there's
34:50
music festivals going on and then all of a sudden
34:53
there's like an earthquake and they end up in
34:55
this like super dome where they have to have
34:57
this epic rock battle. And so it was just such
35:01
a unique. book and story that as I was watching
35:04
the student develop it and I watched the student's
35:07
excitement for school and this is a friend who
35:10
was not always excited about all of the things
35:12
that were happening. It was just, it was amazing.
35:16
And to see this student even like connect with
35:18
their dad over it. And they were talking about
35:20
like how the book developed and like how awesome
35:23
it would be. We did find in fifth grade, especially
35:25
a lot of fan fiction, which is not a genre that
35:28
typically exists for students at this age in
35:31
the elementary school. So we have students dancing
35:34
on stage with Michael Jackson. We have the student
35:36
who ends up in the WWE or gets drafted for the
35:39
Patriots. So lots of unique. really personalized
35:46
books that other students are even connecting
35:50
to. The one you just shared, that honestly sounds
35:53
like a Jack Black script. Yes. It's so interesting.
35:58
So it was the first line of the prologue. I read
36:01
that when it was being generated, and I heard
36:03
Jack Black's voice in my mind, like, listen up,
36:07
young rocker. That's exactly what was happening
36:11
when I was listening to you explain it. I think
36:14
the cover art that's there looks really amazing.
36:18
Like I'm sure that wasn't the first generation
36:20
of an image most of the time that they had to
36:22
keep working on it, right? But man, they look
36:25
awesome. Yeah, they did. They did such a great
36:27
job and it's so interesting because this like
36:31
having adult learners and student learners at
36:34
the same time in this project. The different
36:37
groups struggled with different things and excelled
36:39
at different things and the image generation
36:42
was something that the students really excelled
36:44
with. prompting is something I think that prompting
36:49
in general teachers are really uniquely qualified
36:52
to be excellent prompt engineers because we're
36:54
used to explaining really complex processes very
36:59
very simply and like and we can articulate those
37:01
thoughts but when it came to the image prompting
37:04
the students were much able better to articulate
37:07
the style and the elements and and exactly exactly
37:10
the action that they wanted to see in the images.
37:14
And just to remind people here really quick that
37:16
if you want to see some of these images and you
37:19
are on either Apple or Spotify, that you need
37:22
to go over to the YouTube channel and check out
37:25
some of these different ones that are on here.
37:27
They're really amazing. So I think did I answer,
37:38
I forgot what the original question was. Yeah,
37:40
you definitely showed many of the unique and
37:42
creative ones. And definitely there's an opportunity
37:46
for people to look for more. Yeah, and it's great
37:50
to see like this is kind of some of the end product,
37:53
right? I think there's a part two that is important
37:57
to mention is obviously there through this process.
37:59
And we talked about struggle a little bit earlier
38:01
just in regard to you creating this for teachers.
38:04
We know that students can often face struggles
38:07
too, right? There's challenges, there's frustrations
38:09
along the way. How was that handled with, you
38:14
know, as this process went on and you kind of
38:16
saw students creating these books, what were
38:20
the challenges or frustrations that you witnessed
38:23
and what other valuable lessons could be taught
38:27
there as far as like resiliency, things like
38:29
that? Yeah, by the time we were towards the end
38:33
of this after state testing, we were on a bit
38:36
of a tight timeline. But I think because of the
38:39
novelty of the project and because of its like
38:42
intense personalization, student engagement was
38:46
actually not an issue. So and we had even so
38:50
when they got their first five. book synopsis
38:53
there was also a doc that was filled with like
38:55
other ideas like with like 40 just other general
38:58
ideas and we did have some students who are like
39:01
I don't like any of these like we know we have
39:03
those like okay we'll go take a look there and
39:06
I was like oh well I like this one and so then
39:08
they they're always these like backup options
39:11
for for if we had a student that didn't really
39:16
want to move forward especially when we got to
39:18
that outline piece at the end because if they're
39:20
like oh I like this chapter the way it is it's
39:23
like okay well then add in a piece of dialogue
39:25
or add in what is that character wearing like
39:27
add in at least a detail about it so it wasn't
39:30
that like okay it's good okay it's good I want
39:32
to see what's next it really made them intentionally
39:35
work on it the struggle we had was absences so
39:38
so students that were out But again, we were
39:41
so lucky that they were so engaged that we had
39:43
kids at home sick working on, like the only thing
39:46
they did want to work on was their book. So we
39:51
were so lucky and we also had resources from
39:55
Google for Education available to us too as part
39:58
of the Innovator project. So there was that excitement
40:01
factor too. We had some people come and visit
40:03
the school from Google. And so there was a lot
40:07
of excitement around the project. And we had
40:09
a big assembly at the end where the students
40:11
shared parts of their books and were able to
40:15
even read with a buddy. Awesome. What would you
40:20
say would be one of the most surprising or even
40:23
rewarding outcomes from doing the project that
40:26
kind of caught you off guard? I think the most
40:31
rewarding piece of it was seeing the students,
40:35
and I mentioned it a little bit earlier, seeing
40:36
the students connect over their books. I was
40:40
in the hallway. It was the last week of school.
40:42
I was in the hallway with one of my friends who
40:44
takes walks in the hallway. And I always ask
40:46
for the 411 what's going on in the building.
40:49
And we were walking for a little bit when another
40:51
student went to fill up their water bottle. And
40:54
I went to that student. I was like, oh, your
40:56
book is like the banana blitz. It's this very
40:59
silly book. And I was like, you haven't met this
41:01
student yet. There were two grade levels apart.
41:04
His book is all about this virus that makes objects
41:08
come to life. And they were like, what? I want
41:11
to read your banana blitz book. And the two of
41:13
them literally just walked away from me talking.
41:16
It was the sweetest thing, because I talked to
41:19
both of these students all the time, but they
41:20
had never even really met each other. And they
41:23
just walked away talking about their silly, very
41:26
silly books. And that moment for me was just
41:29
like, it was just the best to see them so excited
41:32
about their idea becoming something that they
41:35
could actually have a full conversation about.
41:37
And it's not just an idea. It's actually like,
41:39
oh, and then this happened and then that happened.
41:42
It was, I think that was definitely the most
41:44
rewarding piece, but also it was surprising to
41:47
see. You think these kids have such niche interests
41:50
and they do, but they're also so interested in
41:53
learning. It's that windows and mirrors that
41:55
we hear a lot about representation. And to really
41:59
see that in action in such a personal way was
42:02
just amazing. And the culmination of the project,
42:09
the students, you just mentioned a couple there
42:11
where they're sharing what they made and things
42:14
like that. There's also the physical copies too.
42:18
And so extra copies get donated to the school
42:21
library, right? So you have, it's kind of a unique
42:26
piece of it because it's not just digitized then.
42:29
You actually have the physical books, something
42:31
that you can, you know, tangible. What kind of
42:34
impact do you think that has on maybe, you know,
42:38
getting students back connected with the physical
42:41
reality of their creation? I think having the
42:45
most of our work was actually done originally
42:48
in Google Docs. But then once we actually created
42:50
that first text, nothing was done on the Chromebooks.
42:53
Nothing was done digitally. They got a first
42:54
manuscript copy that was like on copy paper,
42:57
but it was two pages to. to a page and so it's
43:01
stapled on the side. So it looked like a real
43:02
like manuscript. And so they went through and
43:06
that's where they had to like do the hallucination
43:08
hunt. And that's where they were like doing all
43:10
of their final editing. We found some really
43:13
funny hallucinations along the way which even
43:15
made the students more excited to read their
43:17
books really closely. They did their final editing
43:20
and all of that was done on paper. And so they
43:24
have these manuscripts of these books that are
43:26
totally marked up by then. And then we even showed
43:29
them like examples of I think I forget the book.
43:32
It was it was a book that they had been reading,
43:34
but we found like and marked up editors like
43:37
version of it, a picture of it. And so the students
43:39
were able to see like this is what authors do
43:41
and editors do is they take these these manuscripts
43:44
and they mark them up. And then that's what goes
43:46
in. So they've made these final copies, these
43:50
digital versions that then we got printed into
43:53
actual books. But now they have the both iterations
43:56
of it physically, too. which is really exciting
43:59
for them. And it was such a big deal because
44:01
they all wanted to bring them home. And that
44:03
was part of it, too, is like, no, you can't bring
44:05
it home until the end. But we didn't even let
44:09
them keep them in their desks to build excitement.
44:11
If you weren't working on it, you didn't have
44:13
it. So anytime they weren't working on it, the
44:15
teacher collected all the manuscripts in a bin
44:18
and they had to go and get it in order. It was
44:22
about building that excitement and making it
44:23
special. We didn't want it shoved in the desk
44:25
and all wrinkled up. Now, I can tell this is
44:29
not just a one -off project. It wasn't something
44:31
that you just did for this school year. So, as
44:35
we're gearing up to start another year with our
44:38
kids, what do you see changing? What do you see
44:42
improving? What do you see staying the same?
44:45
Or do you have new surprises up your sleeve?
44:48
I think even just this summer, like I said, the
44:51
models have improved so much. I need to go back
44:54
and take a look at the gems that I had. at the
44:57
end of June. And I think we can streamline now
45:02
so much of the actual process where I think the
45:05
prompts, like you said, when you iterate a gem
45:07
or when you're building your own bot and you
45:10
actually use it to improve itself. I think I've
45:13
gotten to a point with those now that I can really
45:16
train the teachers on how this is working and
45:18
why this is working and get more technical on
45:20
the teacher side. Whereas last year was much
45:22
more proof of concept and being like, OK, this
45:25
is something we can do and we can do it safely.
45:27
But now it's going to be more teaching, training
45:29
the teachers on what's actually happening. And
45:32
I think from the student side of things, just
45:34
our capabilities, like I said, to be able to
45:37
add really amazing images throughout the whole
45:39
book. maybe to record some audio books and we
45:44
want to start at different points in the year
45:46
at different grades so we can build up our library
45:49
and make it more robust at different grade levels
45:52
as well. So there's a lot coming this year. I'm
45:55
really excited to see where the project goes
45:57
and we're going to keep everything on this website
45:59
so as we build these new books they'll be added
46:01
to this same site. Well, I hope you'll let us
46:05
know how the process is going and keep us updated
46:07
on new things that you're doing. But yeah, David,
46:12
you got the last one here? Yeah, Liz. So obviously
46:16
we'll share the website as well in our show notes
46:20
and things. But is there other ways that, you
46:23
know, if people have questions that you're willing
46:25
to engage with, like through email, social media,
46:29
any of those type of things. And then what do
46:30
you have coming up? Do you have any upcoming
46:34
events, projects that you want to promote or
46:37
share? Absolutely. So I'll give you my email
46:40
address also to include. I'm not on social, but
46:43
please anyone feel free to reach out with questions.
46:46
I'm happy to share my gems. I'm happy to share
46:49
the process or any information about this, especially
46:53
to get some feedback for improving it for my
46:55
teachers. I am planning on presenting at the
46:58
Masque Conference at Gillette Stadium in October
47:01
about this project and then about just building
47:04
gems in general. It's something that I've done
47:07
a lot of work on and I've actually created some
47:10
template gems that I'll also give you guys links
47:12
to if you want to post those as well. So I'm
47:16
really excited to see where this year goes, but
47:18
I've loved having this opportunity. This has
47:20
been really fun. Yeah, we've really enjoyed it
47:23
too and thank you once again. Thanks for listening
47:29
to Integrate This. Want to take your learning
47:32
further? Head over to our YouTube channel for
47:34
quick tutorials, deep dives, and more EdTech
47:38
tips you can use right away. Don't forget to
47:41
subscribe and catch new podcast episodes every
47:44
Thursday on Apple, Spotify, or wherever you get
47:47
your podcasts. Until next time, keep exploring,
47:52
keep innovating, and keep integrating this.