Exploring CodeHS Tools: Teacher AI Hints and Smart Hints for Students

Hello and welcome everyone. Thank you for joining our webinar. We're very excited to be here with you this afternoon as we explore another CodeHS tool. Today we'll be talking about teacher AI hints and smart hints for students.

So we'll give it just a few more seconds here as folks trickle in. We're so excited again to show off these awesome new tools. I'll go ahead and put a link to our slide deck for today's presentation in the chat just for your reference if you want to come back to this presentation at any time this school year. When you're ready to try out our teacher AI hints or our smart hints for students, you can come back and reference everything we'll talk about today.

Thank you for being here. My name is Stephanie Bennett. I'm a PD specialist with CodeHS and I've been with the company for a little over two years now, but before that, I was a classroom teacher based in Indianapolis, Indiana. I'm excited to be joined by my colleague Danielle Santilli, and she'll be helping out in the chat and in the Q&A if y'all have any questions about the tools as we show them off today.

Just a quick reminder that live attendance is required for a CodeHS certificate of completion, and we do ask that if you have invited your AI notetaker, please keep that muted and disabled from accessing the chat. You can ask questions in our chat feature in Zoom today or you can ask questions in our Q&A feature. The benefit of asking questions in the Q&A is that your questions won't get lost as we pop links to knowledge base articles and other good information to have in the chat. So if you could put questions in the Q&A, but feel free to add your tips, tricks, and experiences to the chat as we move along.

You will need a CodeHS account in order to access these tools, so we will ask you to sign up for one if you don't already have a CodeHS account. They are free for teachers and students to access our fantastic curriculum. Lots of high quality instructional materials are available for teaching many different computer science subjects, and so we love to provide that curriculum free to teachers to use in their classrooms.

Part of our mission at CodeHS is to ensure that all students can meaningfully impact the future. To that end, we are a comprehensive K through 12 curriculum for use in elementary school, middle school, and high school to bring high quality learning experiences in computer science subject areas to your students. Everything is web-based, so we find that it's an equity driver. Keeping all of our tools web-based means that any student that has a device connected to the internet can participate in computer science learning. We also offer PD, which you know since you are here with us today, and we also have a suite of LMS tools that you might be familiar with from using other edtech products. We provide the ability for teachers to give feedback to students on their learning, grading tools, and the ability to send messages to students or announcements to whole class sections. Lots of those tools that you may be familiar with, we provide free to teachers.

Welcome again as you are getting settled in. We are here to talk about some of our newest AI products. Before we dive into the specifics on our teacher AI hints and our smart hints for students, we do want to share our approach to AI. Our goal today isn't to tell you that you should be using AI in your classroom. We want to show you rather that we have these two tools available within the CodeHS platform, how they're designed to support teaching and learning, and also make clear the controls that teachers and schools have over how these tools are used. As you see these today, we will ask you to think about whether and where they might be useful in your classroom.

Before we look at the tools, there are three things worth understanding about how CodeHS is approaching AI usage for teachers and students. Our approach is that districts get to decide which features are available based on their own policies. Within those settings, teachers have flexibility over what fits in their classroom, their teaching style, and what's best for their students. The tools themselves are designed specifically for an educational environment with protections and controls built into the experience. If you'd like to learn more, thank you Danielle for dropping that link in the chat, we do have the CodeHS AI Center that helps explain how we protect student data, teacher data, and how our AI tools are used behind the scenes.

Let's jump into these exciting two tools that we have. The first one is our teacher AI hints. I have a bit of a challenge for those of you in attendance today. I just want to give you a visual of why AI can be helpful in a computer science classroom and specific applications for AI tools. The first one we're going to do is a debugging challenge. We're going to navigate to this sandbox program and you'll be able to fork a copy for yourself. I would love for you to grab your cell phone and open up your timer. I'll give you a second to get ready for this little challenge. You need your program open on the CodeHS platform in the sandbox and you need a timer, probably handy on your cell phone.

Oh my gosh, did I not give myself permission to view this program? Danielle, can you check and see? I was able to get into it. Okay, interesting. Well, lucky for me, I have it here in a section.

All right, so we're going to time ourselves debugging this JavaScript program. Imagine you're in class, a student raises their hand, you come over to their device, and they've got this program pulled up. They run it for you, and you can see that there are some errors. How long would it take you to debug this program? Go ahead and start your timer and find the bugs.

All right, so maybe some of you have finished up now. For me, I would have to remind myself, okay, what class period is it? What language am I teaching this class period? Okay, this time it's JavaScript. That means that we definitely can notice some syntax errors here. For example, this semicolon needs to move outside of the closed parenthesis. Here we are using an assignment operator when we should be using a comparison operator. Let me run it now. Nope, we still have an error. Okay, so here we are asking the user for a boolean. Nope, we are looking for a grade level, so that would be `readInt`. Let me run it now. No, we've still got something wrong here. Oh, here, our `if` statement is not enclosed in parenthesis. So let's see here now. Okay, now we got the enter the grade level dialogue box. Oh, but we've got an invalid answer. So you can see that this is taking me several minutes to figure out all of these bugs.

If I were to instead entrust this to a chatbot LLM, let me go back and reintroduce some of these bugs that I fixed. If I copy and paste this code now and tee it up to, let me say ChatGPT, let me see how long it takes ChatGPT to find the bugs. This is hilarious, Danielle. It's going to take as long as I did. I'm not logged into my ChatGPT, this is just using it for free, so here we go. It's processing. We're using an old model here. ChatGPT is taking too long; I'm going to Gemini. Yeah, maybe it's down or maybe it's my internet.

All right, so here Gemini at least is working for us today. It gives us some of those items that I identified myself, but you can see at least with Gemini, maybe not GPT today, we get corrected code really, really quickly. We can double-check that this does indeed work by copying and pasting it back into our program. Let me run it here. I'll put in grade nine, and we get the response, "wear a gray t-shirt."

Gemini did that in a fraction of the time it took me, the teacher, to debug one student's program. If you can imagine, as nervous as I was when ChatGPT was failing me there, starting to sweat, the same thing happens in the classroom when you have four or five students' hands in the air. Everybody's stalled because they're not sure what's wrong with their program, and the pressure is on for you, the teacher, to find the correct answer, but not give it away to the student, right? Like that's the other thing. We don't want to just give the answer to the student; we want to coach them to the correct answer. That is a lot of processing to do over and over and over every single day. Gemini took fractions of the time it took us to find the bugs, so we can tell that AI can be a very valuable teaching tool for us in the classroom.

CodeHS has built AI teacher hints, and the tool allows teachers to debug student code in real time, provide coaching and feedback to students, help students think more deeply about their coded responses, and quickly identify a student's next step without giving away the answer. Great functionality here with this tool.

I'm going to hop over again to my platform and get into a course where I can show off this fantastic tool. If I go into any student's work, let me hop into Ada's work here, we can look at a program that the student is working on. If she has a question about the program, we can move to the Conversations tab, which lives under the More menu item in the editor here. This is where students type their code in the CodeHS platform, and we can open up this Conversations tool. Here at the bottom of this message box, I have the ability to generate an AI hint for Ada. This program is finalized and it's great, but it looks like we do have some advice for Ada when we use that Generate AI Hint option. You can see here also that it's totally editable by me, the teacher, so that I can make sure that this response is appropriate for this student and actually helps them without giving away the answer. When I'm happy with that message, I can just click send, and now Ada in this assignment can see that message on her end when she is working.

AI hints are a great tool, and you might have noticed that we actually have three different levels of hints that we can provide for students. Let's dive into these options and talk a little bit more about the usability and the nuance to this teacher tool.

AI teacher hints help us generate feedback based on a student's current snapshot of their work. AI hints are only available for teachers to generate, and you could see in that workflow that we keep the human in the loop. Hints have to be reviewed before they're sent to students, and all hints are fully editable.

Let's talk about these AI teacher hint levels. General hints are going to be something like, "don't forget to read the directions," right? We know students sometimes dive right in without even looking at an assignment. In CodeHS assignments, sometimes there's a very particular variable name or function name that students are required to use in order to get their code to pass the autograder. General hints are going to be pretty vague, something like, "look at the directions again," or "note in the directions that you're supposed to use this boolean."

Identifying the problem gives students another layer of support. Going back to our example, "it looks like you didn't use the correct variable name." Again, we're not giving away the answer to students with these teacher AI hints; it's just very clearly identifying a problem. It might even say something like, "it looks like on line 12 you didn't use the correct variable name." Identify the problem makes it very clear to students where their code is not matching up with the canonical solution code.

Identify the problem and offer a solution is again just one additional scaffold for the student. If we go back to our example of the variable name, "it looks like you didn't use the correct variable name. The directions say to use a variable named `year` in your code." Again, not giving away a full coded answer, but redirecting the student's attention to the directions and the requirements of the assignment.

That is our AI teacher hints. I just want you to notice that you're not just clicking an AI button and accepting whatever comes back. We're leaving it to your professional judgment for you to choose the level of support you want to provide to students because maybe a student only needs a gentle nudge to reread the directions. We wouldn't want to coach them too strongly if all they're doing is hitting the easy button for the day. But another student might benefit more from having the problem identified clearly to them. Here we're keeping the power in the hands of the teacher to be that instructional coach to their students.

Danielle has put a link to our AI teacher hints Knowledge Base article in the chat. We encourage you to check out that Knowledge Base in the future if you are looking to try out AI hints and want to learn more about it.

To use these, again we'll just work through that workflow in any student assignment. You can open up the assignment, click on the More menu option, and find the Conversations tool. You'll then scroll down, click Generate AI Hint, and select the type of hint that you want to provide to the student. Finally, you'll review that hint, that generated text, make any edits you see fit, click Send, and then we would love it if you would give us some feedback, thumbs up or thumbs down, really quick, on how the AI hint turned out. Was it useful for your students? Did you appreciate the assistance and coaching it provided? We do take these responses and this feedback very seriously and use it to make our AI tools better. So that feedback to us about the quality of AI hints is very useful just in the quick form of a quick thumbs up or thumbs down.

All right, well let's move on to smart hints for students. This is one of our most recent tools that we have been developing here at CodeHS. Smart hints are a student-facing AI tool. However, it is not a text-based free response text box interaction that students can have with our AI Carol. Instead, students are provided with four options for asking AI Carol for help. You can see these here on the screen.

This isn't designed as a general purpose AI chatbot. Students aren't sitting at an empty text box trying to decide what to ask. Smart Hints provides these four specific options and only these options tied to the kinds of problems that students commonly encounter when they are coding. Just to reiterate, because we do want you to have a clear understanding of this student-facing tool since it is provided to students if you want it, we want to be very clear on what it is, what it isn't, how it helps students, and benefits students. Again, students cannot type freeform questions into this tool; they can only select from the four prompts. There is no ability for them to expand on this chat at all.

Something really good to know is that the tool also is limited. If we do decide to turn smart hints on for students, then we would need to help students understand that they only get five hints per day. They have to be judicious about when they're choosing to turn in their AI hints. I think that's a really great teachable moment for students on the limitations of AI and using it judiciously and responsibly.

When you turn on AI hints, you have full visibility into students' usage of the tool. You can see how many times they use it, on what assignments they're using it, and you can also see which hints they asked for. We provide this level of visibility and monitoring because we think not only is it great to keep these conversations going with students about their AI usage and potential moment of relying on an AI chatbot a little too much, but we provide this data to teachers because students hitting five out of five hint limits per day is definitely going to be a strong signal that we should check in with that student. We should see where their misconceptions are. Is it just a problem with getting started with the code? Well, then what other resources do students have besides the AI hints? Is it because they don't remember how to write a `for` loop and they just keep asking what's the syntax for a JavaScript `for` loop? Then what other tools could we provide, what other scaffolds could we provide to that student that would avoid the usage of the AI hint? You do have insight into all of that data through our student usage dashboard, and you have the ability, like I said, to see exactly which hints students were using through that chat history.

Let's review our two tools for today: our AI teacher hints and our student smart hints. We are providing these tools as a support, but we want to reiterate that teachers remain in control of their classroom and every student's learning experience. As you think about AI in your classroom, these tools aren't intended to replace your teacher feedback, and they are definitely not intended to replace student problem solving. They're just additional supports that schools and teachers can choose to use when and where they make sense. These controls we've looked at today, from the teacher review and editing in the AI teacher hints to these preset prompts that students have available in smart hints, all of these are locked in as part of the tools themselves so that you don't have to build these tightly controlled AI tools from scratch.

We'll just end today with one more view of our AI Center where you can learn more about the CodeHS approach to our AI tools. We know that AI is changing how schools teach computer science, and our approach is to make sure that usage of AI on CodeHS supports your school and your district's learning goals and gives you the control with how it's implemented.

We definitely appreciate your time here today as we previewed these two tools. If you would like to stay in touch, you can join us on social media if you have those accounts, stay in touch through our certified educator program, or even earn micro-credentials with us. You can learn more at codehs.com/free. We also invite you to check out our new webinar library. If there is a certain coding language or course that you're wanting to learn more about or previous CodeHS tools that we have been focusing on, find those at our library.

Of course, we will provide a certificate of completion to you. We don't do links for those anymore; instead, we automatically upload the Zoom attendance to the CodeHS site, so those will be generated automatically. Please just allow 48 hours for receipt of your certificate before you reach out with any questions about that.

We thank you so much for joining us here today. The last thing we'll ask you to do is give us some feedback on how we did in our webinar survey. We do take our feedback very seriously here at CodeHS. We're always wanting to know how we can serve teachers better and provide great tools and curricular resources for you in the classroom, so definitely give us a little feedback on our survey. Then we'll just say thank you, and if you are interested in learning more about bringing any CodeHS Pro tools to your school or district, you can find out more with the link in the chat. Thank you all so much for your time today.