#422 – CS Education in the Age of AI, with Perry Shank

How do we prepare students for a future shaped by artificial intelligence without losing sight of the human element? In this episode, Perry Shank, Senior Director of Research and Development at CodeVA, answers that question while exploring how computer science (CS) education is evolving in an AI-driven world. He shares his personal journey, touches upon the importance of understanding AI—not just as users but as informed creators and collaborators—and offers powerful insights around integrating AI meaningfully and ethically into the classroom.

We prioritize quality computer science education for all students and promote the connection of CS education to economic development.

CodeVA, from the About Us page of their website

Resources

The following resources are available from AVID and on AVID Open Access to explore related topics in more depth:

• AI in the K–12 Classroom (article collection)
• An AI Literacy Framework (podcast episode)
• AI Trust You (podcast episode)
• SchoolAI, with Kevin Morrill (podcast episode)
• The A.I. Roadmap: Human Learning in the Age of Smart Machines, with Dr. John Spencer (podcast episode)
• The AI Infused Classroom, with Holly Clark (podcast episode)
• Academic Integrity in the Age of AI (podcast episode)

Humans in the Loop

As artificial intelligence becomes increasingly embedded into daily life, computer science education must evolve to help students navigate and shape this new terrain. At the heart of this transformation is a powerful idea: keeping humans in the loop. Perry Shank emphasizes that while AI tools can be powerful collaborators, they should never replace human reasoning, creativity, or ethical judgment. “We have to make sure students are not just asking for an output and running with it,” Perry says. “They need to balance machine reasoning with their own sense of judgment.”

This episode explores how educators can foster that balance by demystifying how AI works, spotlighting human decision-making in data and design, and engaging students in thoughtful use of these tools. From hands-on activities—like Google’s Teachable Machine and micro:bit CreateAI—to building custom Generative Pre-trained Transformers (GPTs) for critical conversation, Perry highlights the importance of curiosity, experience, and reflection. AI literacy must be as foundational as math, and it must center the learner as a thinking, ethical, and empowered human in the loop. Here are a few highlights from this episode:

• About Our Guest: Perry Shank is the Senior Director of Research and Development at CodeVA. Perry began his career as a music teacher and has experience teaching across the K–12 spectrum. He has also taught art and AP computer science. CodeVA was the organization that helped train him as he transitioned into the role of computer science instructor.
• Interconnectedness: Computer science is connected to many other disciplines and should not be studied in a bubble. Perry feels that CS should be approached in a “transdisciplinary way to things that students are learning in other classes or maybe experiences that they might have outside of the classroom.” Speaking about AI, he adds, “It’s embedded in our news. It’s embedded in our lives. It’s embedded in our classrooms. And grappling with artificial intelligence tools and having rich discussions about maybe ethics or the function of a tool and how it’s built, like all of those things, are great learning experiences for CS.”
• A Desire to Learn: Perry says, “They [teachers] want to know about AI a lot more, and there’s a lot of great tools out there now and a lot of resources for them to experience that.” He adds that when ChatGPT was introduced, “The interest that teachers had in AI really exploded.”
• Responding to the Demand: In response to the demand, CodeVA began developing resources to help teachers. Perry explains, “We jumped into it, and we started building out some products about teaching, like: What’s in this black box of AI tools, and how are they formed? How do they work? What are their use cases? What are some of the barriers that are presented?”
• Going Beyond Using a Tool: Perry shares, “I feel like a lot of teachers are really interested in learning more about how to engage students in a very impactful way about AI content, beyond just use of AI as a tool, but actually understanding why it gives you the results it gives you.”
• Understanding Intent: Perry says, “We have to make sure that we’re not forgetting that these tools were created by someone, right? And they’re designed in a very specific way, and so understanding the way that it’s designed is going to help us learn how to use it better.”
• Human Reasoning: When using AI, Perry believes that students need to be critical users. He says they must learn to “figure out its shortfalls, to figure out what it does really, really well, to figure out what the human interaction is within the use of an AI tool.” He adds, “I think we really have to be aware and talk to our students about making sure that they’re not just asking for an output and then taking that output and just running with it. They have to think about the balance between their own human reasoning skills and machine reasoning, how they use their own sense of judgment, and the way that they make decisions as a human as they use the AI tool.”
• A Subtopic of CS: Perry sees AI as a subtopic within the computer science field. In fact, he says that labeling AI is not extremely important because the various aspects of computer science are all interconnected and interdisciplinary in nature. He says, “We actually need to teach students all of those things and how they connect together.”
• Google’s Teachable Machine: A useful tool for helping students understand how datasets shape machine learning and AI functionality is Google’s Teachable Machine. By using a webcam and microphone, students can input datasets into Google’s Teachable Machine and learn how the specific data being added impacts the functionality and performance of the machine. “The tool is really cool because it shows you the confidence level that the training set has in what you’re offering it through the webcam,” says Perry. “I’ve seen teachers do amazing things with this tool and also just have this light bulb flash above their head, of like, ‘I get why this is tricky.’”
• Another Relevant Tool: Perry adds, “I’m a big, big fan of micro:bit and just thinking about all of the different sensors that are on board that little device and how you might use that as data input for your training structure. It’s really excellent.”
• RAG Setup: Perry talks about the value of limiting the data that AI can access through the use of a Retrieval Augmented Generation (RAG) setup. RAG involves identifying and limiting the dataset that a large language model will use to answer questions.
• Feeling Overwhelmed: Perry acknowledges that it’s easy to become overwhelmed with the speed and volume of information regarding CS and AI. He feels that it’s important for teachers to have a community of trusted resources upon which they can rely to process the new information. His second tip is to choose a tool and try it out: “There’s nothing that teaches you like experience.” It’s also okay to struggle and fail along the way. He adds, “We learn a lot from failure.”
• Empowering Staff: According to Perry, “One of the best things that a school district and a school can do is actually train their staff, or send their staff to a training where they can learn new things and then bring it back to the school.” This empowers local staff and ensures that the new learning is done in the context of the local district. Perry says, “The best way for teachers to learn is from their peers. And I think that prioritizing building their own human capital is the best way to do it because they know their context.”
• Human in the Loop: “I’m really interested in exploring what it means to be a human in the loop and to devise a process for that within an AI tooling structure,” Perry shares. He adds, “I’m really excited in thinking about how learners—both teachers and students—can approach the use of AI with this ‘human in the loop’ perspective so that we don’t lose the human qualities that make learning exciting and that make working with each other beneficial.”
• MegaBytes: MegaBytes is a student magazine created by CodeVA, and it includes a feature called Artificial Imposter. “In every issue, the artificial imposter tries to trick the viewer into thinking that something’s real that’s not.” Perry explains, “It’s great for discussion and for sharing with your family.”
• Artificial Intelligence Pathway: AI Pathway is a set of curriculum that can be used for purposes such as a high school programming course. The resource will be launched soon by CodeVA and addresses the topic of designing AI structures. Perry points out that it’s about “actually going beyond just using AI, but designing AI and learning deeply about AI through that type of exploration.” If you would like to learn more prior to the product launch, a related interest form is available.
• One Thing: For his one thing, Perry talks about the importance of designing structures and learning opportunities “for all students.” He adds, “We’re talking about engaging with AI where we’re deepening students’ ways of thinking critically about the world, around designing in ethical ways, and really staying grounded in those human values.”