How to Preserve Humanity in Education in an Age of AI

Chief Learning Officer Dr. Tim Hudson with highlights, caveats and insight.

INTERVIEW | by Victor Rivero

Dr. Tim Hudson serves as Chief Learning Officer at Discovery Education, where he supports the research and development of digital, personalized lessons and assessments in math and literacy that engage students in critical thinking while also providing teachers and administrators with insights to support differentiation and improve student achievement. Before joining DreamBox, Tim spent over 10 years in public education, first as a high school math teacher, and then as a K–12 district mathematics coordinator and strategic planning facilitator in suburban St. Louis, MO.

I recently sat down with Dr. Hudson to talk about a topic that is top-of-mind for educators worldwide: Artificial Intelligence. Read on to learn Dr. Hudson’s thoughts on the benefits for students and teachers, what educators should look for as AI continues to permeate education, and how to preserve “humanity” in education in the Age of AI. 

Since the rise of ChatGPT last year, Artificial Intelligence has been on the minds of educators worldwide. When AI burst on the scene, it seemed as if the focus was primarily negative. Play the contrarian here—as AI continues to push into education, what are the benefits for teachers and students?

I think initial responses about the educational applications of new AI tools were mostly influenced by a person’s understanding of today’s classroom realities. As with other technologies in the past – radio, TV, PCs, the Internet – there’s been plenty of irrational exuberance about how AI will “transform education forever” because ChatGPT can pass challenging written exams. That’s impressive for sure, but when we consider the benefits, I always note that they are not as revolutionary as some claim. 

As a math teacher, this moment with AI reminds me of when calculators – and later Wolfram Alpha – came onto the scene. Because calculators do tedious number crunching and graphing, math teachers were freed to focus on big ideas, assess deeper understanding, and develop critical thinking so students can validate whether the calculator output is reasonable. And I think that’s also the biggest educational benefit for new AI tools – ChatGPT can do tedious basic writing and help teachers generate discussion prompts, which frees teachers to focus on developing critical thinking skills that equip students to analyze the validity of writing regardless of whether it’s from AI or other people. Like all technology tools, AI should be used to cultivate and improve critical thinking.

‘Like all technology tools, AI should be used to cultivate and improve critical thinking.’

But the implications for AI in the classroom go beyond the delivery of instruction, right? AI can sift through massive amounts of data and spot patterns and relationships and help policymakers and other education leaders make new insights into teaching and learning. Can you share with readers what that might look like, and how the benefits of that ability will be seen across the education spectrum?

I believe there are not yet enough large, meaningful, education-specific data sets that AI could analyze to help us tackle our most persistent and challenging problems. As one example, my doctoral research focused on new teacher job satisfaction because even 15 years ago, 50% of teachers were leaving the classroom within their first five years. I wrongly assumed that such a pressing problem would have been researched extensively. But there were very few studies, and none had large enough data sets that would require AI to analyze. 

In places where we do have large, longitudinal data sets – such as NAEP (National Assessment of Educational Progress) scores and demographic information – we don’t yet have data at scale about what schools are doing to address problems like the ongoing opportunity and achievement gaps for Black and Latino students and students experiencing poverty. So, AI could only surface things we already know, such as how socioeconomic success factors are highly correlated with student achievement. That’s one reason why our partners in East Baton Rouge Parish School System were awarded a $3.5M Gates Foundation grant to work with us and Louisiana State University as we explore ways to gather and analyze new data sets. 

DreamBox Learning, which was recently acquired by Discovery Education, uses AI in its products to create customized learning paths. In his June 6, 2023 interview with me, Discovery Education CEO Jeremy Cowdrey mentioned that the company’s other solutions use AI, and that the plans for future uses of AI will be viewed through the lens of how it helps the teacher and supports instruction. Can you put your futurist cap on for a moment and supply some insight into where AI is not being used now that it could someday be used to support teaching and learning five years from now?

Any technology or tool is a means to an end. So, when thinking ahead to how AI could be used five years from now to support teaching and learning, we need to define what we’re hoping to achieve. 

‘…when thinking ahead to how AI could be used five years from now to support teaching and learning, we need to define what we’re hoping to achieve.’

Building on Jeremy’s point, at Discovery Education, we care deeply about making sure teachers and students have quick access to the most engaging and effective learning resources. That’s because our goal is ensuring high achievement for all students while closing opportunity and achievement gaps. So, we’ll be exploring uses of AI that help teachers and students find and connect with the most relevant and appropriate content and learning experiences every day. We see ways to help teachers more quickly find and use the perfect Discovery resources each week. We’ll use AI alongside the extensive internal data we have and use the wisdom – the human intelligence – of our teacher and district partners to create exciting new possibilities such as more strategically scheduling their time and student strategy groups in ways that make learning more fun and engaging. 

So far, we’ve painted a positive picture of the promise of Artificial Intelligence in K-12 education. Can you speak to some of the “look-fors” in the use of this technology? What does the education community need to keep in mind as this new technology is deployed?

I’d suggest educators ask themselves at least these two reflection questions as they navigate our new reality with AI:

1. Are you using AI as a means or as an end? If your goal is simply to use AI because it’s new and flashy, then you’re unlikely to achieve your goals for improving student achievement. AI is a tool, and like all tools, it carries imperfections. So, I encourage educators exploring the uses of AI in the classroom to begin by defining your goals first and then deciding to what extent AI could be the “right” tool for the job.

2. Are your assessments “AI-proof” and “calculator-proof?” I’ve said for over decade when presenting at math conferences that if Wolfram Alpha can get an A on your math tests, then now is a great time to develop better tests that require more critical thinking from students. The same can now be said of any non-math written assessments. If ChatGPT can get an A on them, you need to redesign them to make them AI-proof. Teachers will appreciate having better tests because there’s a reduced risk of cheating and plagiarism. Students will appreciate it because they aren’t being asked to do something that AI can do.

You mention the need to keep the “humanity” in teaching and learning in the age of AI.  Can you expand on what you mean there, and give some concrete examples of what that looks like in the classroom?

AI built using Large Language Models can’t be original or independent. It’s limited by the texts it’s been trained in. If we don’t know what texts the AI has read, or if the AI has been fed a steady diet of problematic texts, AI’s lack of original and independent thought poses grave risks. In contrast, our students can read the same problematic texts as AI, but because they are capable of original and independent thought, they will bring unique insights and novel perspectives into the classroom with them they will help them navigate those difficult texts and hopefully identify the biases or flaws.

“Keeping the humanity” means we create learning spaces, experiences, lessons, and assessments that invite, honor, and celebrate each student’s unique intuition and ideas. One concrete way to do that is for students to make connections between historical events or literature and their own lives and interests. Wikipedia or ChatGPT have limitless information about any historical figure or character in literature, but neither of those resources can make a novel connection between those figures and a student’s family member. I’d love to see students writing essays titled, “Three Ways My Great-Aunt is Like Harriet Tubman” because that’s an essay celebrating humanity that AI can’t write.

‘I’d love to see students writing essays titled, “Three Ways My Great-Aunt is Like Harriet Tubman” because that’s an essay celebrating humanity that AI can’t write.’

What’s your prediction for the state of AI in education? What will the use of AI look like next year? How about five years from now?

I think we’ll be talking much less about AI five years from now than we are this year. Already some industry groups like CCS Insight have predicted generative AI has been overhyped and will not command as much attention in 2024. I think we’ll only still be talking about AI in education like we are today if it can easily be connected to improving student outcomes and engagement and teacher job satisfaction and retention. 

As we discussed earlier, even though we leverage AI in several ways at Discovery Education, that’s not the main thing teachers, students, and administrators mention when they describe what they love about partnering with us. Five years from now, they’ll still be looking for relevant, engaging, and effective learning experiences that cultivate critical thinking and are easy to use. So, we’ll only be talking about AI in 2028 to the extent it can help us provide educational content that achieves those goals.

Victor Rivero is the Editor-in-Chief of EdTech Digest. Write to:


    Leave a Comment

    %d bloggers like this: