Getting involved in his school district’s discussions about AI was a no-brainer for Ricardo Vela.

The veteran middle school history teacher doesn’t fit the common stereotypes of an educator who is enthusiastic about new technology — he’s not an early-career teacher fresh out of college, and he doesn’t teach coding full time or serve as the district’s top tech leader.

But that’s exactly why Vela is playing an important role in the rollout of generative AI in the White Plains City Schools, a 7,000-student district in New York.

The educator, who also helps other teachers integrate tech tools into their classrooms as a “computer lead teacher,” brings experience to the implementation process, as well as curiosity.

For teachers in school districts who jumped with both feet into exploring generative artificial intelligence, this year’s national ISTE+ASCD conference in San Antonio — where AI products were everywhere — marked another point in the evolution of the technology in schools.

Early adopters’ experiences are creating a blueprint for what implementation will look like in schools and districts as more systems purchase AI-powered products and tools. Their observations also offer insight into the common hurdles educators will face in bringing a fast-adapting technology into the classroom.

EdWeek Market Brief Staff Writer Emma Kate Fittes sat down with Vela during the ISTE+ASCD event last week to talk about how his district is using AI, what their implementation of AI tools looks like, and how he’s helping the system navigate hurdles — including pushback from some educators.

The following has been edited for length and clarity.

How did you become involved in your district’s discussions about AI tools?

I’m what’s called a “computer lead teacher.” I help teachers integrate technology into their classrooms, and [I’m] also a member of the district technology team.

The year before [last,] we started having conversations related to AI. And we started looking at what tools we wanted to implement in our district.

Members of the team tried out different tools, and it really came down to two for us: MagicSchool AI and SchoolAI. We went with SchoolAI.

This year, we’ve really been building out our AI guidelines for next year.

What prompted the conversation about AI in your district — was it looking more at student-focused or teacher-focused tools?

There was an acknowledgement that kids were using AI without guidance, and they were using it to cheat — let’s be honest.

We realized that we need to get ahead of it, because it’s coming, it’s not going away. It’s not like people can stop using AI, and we would be better off as a district having guidelines and really beginning to teach kids how to use it appropriately. We didn’t shy away from it.

A lot of districts have said, “We’re not touching this yet. We’re not sure where to go with it.” We’re like, “Let’s tackle it.”

It’s better to tackle it now and be proactive about dealing with any issues that might come up.

How did you first start using AI as a teacher?

I really started just by using SchoolAI [and ChatGPT] to write my lessons — not my lesson plans, but my worksheets — and come up with activity ideas.

This year, it has just grown. I use ChatGPT and SchoolAI all the time to write my lessons, to write my units, to design activities, worksheets, and scaffolds for my multi-language learners — everything I can possibly think of. Whenever I have a problem, I ask [AI].

How did your team go about rolling AI out to other teachers and schools?

In each of the secondary schools, we picked several teachers to pilot SchoolAI. Those teachers got access to use it all this year to see if it was really going to be a good fit.

And it certainly has been a good fit. It’s been a wonderful tool to use to lighten my workload, [and] really to just bring learning alive for students.

[The pilot] hasn’t just been a few tech-savvy teachers … but across multiple disciplines: me in social studies, a science teacher, an [English/language arts] teacher.

Have you faced any hesitation or pushback from other teachers or administrators?

My principal has been very open to training teachers in using AI. Teachers are, for the most part, pretty receptive towards it. There are a couple of [people who are wary of new tech], but there’s always going to be [those folks] in every school.

We have a lot of young staff. Because of that, we have a lovely openness and sort of willingness to explore this and see how it can be useful.

What are the next steps of the rollout of SchoolAI at your district?

The rollout is going to be slow. When we first became a 1-to-1 school and everybody got iPads, teachers became totally overwhelmed.

“Here’s this app and this app and this app and this app” — it was way too much.

We made a strategic decision between the principals and the team to do [the rollout] piecemeal. The first group of people to have [full] access to SchoolAI are going to be the ELA and social studies teachers, because that seems to be where it lends itself [as being] the easiest to use.

Then, the second half of the school year, we’re going to bring in the other disciplines.

What about social studies and English/language arts lends itself well to the use of AI tools?

We can still be language-based. There’s a lot of speaking and writing that goes into it, and a lot of creativity. It’s a little harder with math, and a math teacher is a little more hesitant. Math teachers have a hard time envisioning how [to] do linear formulas on a text-based application.

As a teacher, why is it important for you to be involved in this decision-making process for your district?

I’m extraordinarily lucky to work at White Plains. The district committee [that reviews technology consists of] teachers, IT, and teaching assistants who might be computer specialists.

It’s really all teachers, and then the head of instructional educational technology.

What we decide as a team is what [the top ed tech administrator] then brings back to the board and says, ‘This is what we’re going to do.’

Can you give me an example of how you use AI in your classroom now?

I created a bot to help students write an essay. I really created five bots to help them with the different parts of the essay, and it was really cool.

I had one that would help you write the introduction. It wouldn’t write it for you, it just asks you guiding questions to get you thinking about the essay topic and give you help that way.

Because what happens when you start [a class on] writing an essay, if you have 25 kids, 12 of them don’t know how to start. I created this because I can’t help 12 students at once. If I’m trying to do that with one, 11 kids were doing nothing.

With this, the bot will help you if you get stuck, then call me over. What was amazing about it is that I was able to sit back in my half-moon desk and have one-on-one conferences with each kid, and not just talking about [their essay], but [asking] “How are you feeling, and how are you feeling about this process? How’s your day going?”

In that way, it really allows me to build relationships, because I have an assistant that never gets tired, who can help everybody, allowing me to connect with students.

What kind of response did you get from students?

As I was meeting with those kids, one-on-one, I asked them “Are you using it? Is it helpful?” [and] they’re like, “Oh, yeah, it was helpful.”

I presented at our regional [meeting of school districts] on some of the stuff that I’ve been doing with AI, and specifically with SchoolAI, and we brought students.

Afterwards, the teachers, administrators, and superintendents who were there got to interview the students and ask them about their experiences.

It was all very positive. They talked not just about how they used it in my class, but how they used it in other classes to help them solve problems that they’re having or tutor them.

A major concern with AI is that it can be inaccurate or biased in its responses. How do you handle that in your classroom?

That’s something that I’ve given a lot of thought to this year.

Here’s what I’m going to do next year: My whole first week is going to include lessons on the appropriate use of AI.

The three things that I’m going to tell students is, No. 1, that it’s guessing your answer based on everything it was trained on. It looks for trends.

No. 2 is that because it’s just guessing, it can be wrong. The third is that it’s biased because it trained on certain things and not others.

That is a starting point. Then [comes] teaching them, “How do you fact-check that [output]?”

So how do you fact-check?

There are different ways you can do that. You can have [students] find other sources to fact-check what ChatGPT, for example, said. Or you can even ask ChatGPT itself, “Hey, is that really true? Is that accurate?”

Just teaching the kids those skills is really the way I think you can combat that [concern].

What other common challenges have you navigated through in exploring and adopting AI?

When I first started using AI, I was like, “Oh, here’s a tool. It’s gonna give me a worksheet and I’ll prompt it.”

And it wasn’t perfect.

That really took a shift in mindset on my part to say, “Wait a minute, when you use an AI tool, it’s not going to give you something perfect.”

It doesn’t matter what tool you use, you have to prompt and re-prompt and re-prompt, and that way you fine-tune everything.

What’s the most helpful use case for AI you’ve run into so far?

Grading. I uploaded the assignment. I uploaded my documents explaining what needs to be included in each of the body paragraphs of the introduction, and I uploaded the rubric.

I uploaded [the essays] into the chatbot, and I said … “Give me a grade based on the rubric, [and] give me a list of things that they did well, written in my voice, that I can copy and paste with areas where they could use improvement.”

All I had to do was skim [each] essay [and] look at the grade. Did I agree with it? [I] look at the feedback. Is it good? Then [I] tweak the grade.

I was able to grade 100 essays in two days, as opposed to two weeks. That’s huge.