Interview with Dr. Paul Kim, CTO Stanford University Graduate School of Education

Note: Interview was originally excerpted in our recent Impact Report (November 2023) accessible here.

Selected questions and (highly abbreviated) summary answers transcribed below.

Paul, can you tell us about your background and what you are doing today?

I was a computer science major in undergrad and then moved into education technology for my MA and PhD. I've been at Stanford University for the past 23 years developing various edtech tools, and recently developing education games and integrating AI as the Chief Technology Officer and Associate Dean of the Graduate School of Education. In recent years, I’ve been particularly interested in helping students develop critical inquiry skills and creativity. One of those projects I started a decade ago is named SMILE (Stanford Mobile Inquiry-based Learning Environment). I've been working with schools around the world with this particular project, and now I'm integrating generative AI.

You’ve been deeply embedded in the intersection of technology and education for over 20 years. What is your evaluation of our current moment today, particularly with the development of AI? 

Especially through the pandemic period, we've seen a lot of new edtech solutions emerge: from online classrooms to virtual reality, field trips, and science simulations all the way to generative AI-powered tools. In 3-5 years, I believe generative AI will be highly effective in accelerating general productive and personal development. So complex problems that used to take multiple steps and precision coding, will be solved with a simple, rough input. 

I believe as we depend more on AI, students will be asked to use AI tools to solve problems, and therefore students will need to have multiple AI competencies or “meta AI competencies.” The reason that I say meta AI competencies is because in 2-3 years there will be a lot of ChatGPT-like tools, or even more sophisticated AI tools that are available. Educators will start to evaluate students based on their abilities to use these tools and adapt those metrics and rubrics to evaluate student performance with AI. 

What I don't think will happen is AI replacing teachers, or at least, anytime soon. It's quite a contentious topic. However, I do see teachers' roles transforming into becoming coaches. With the coaching model, you have to know the strengths and weaknesses of all your students and help them develop their competencies so that they can reach their fullest potential. And I think that will be the job of the teachers of the future.

How do you think AI will impact the student learning process? What will be possible in the future that is not possible today? 

In classrooms where teachers are not actively integrating AI in teaching, students won't be impacted. Students will remain the same and perform as they have been. But if teachers are actively integrating AI, I think these students will be developing the competencies required to be a productive citizen of the future. That's why I encourage my colleagues and other educator friends to leverage AI and try to integrate AI into classroom tasks. So if teachers provide students with problems that require AI to solve, those students will quickly develop AI competencies. In order for students to be able to be really functional and thrive in the era of artificial intelligence, they will need to develop prompting skills and critical thinking skills. 

I believe speed is the most significant element in productivity. With the help of AI, examples can be much more quickly generated, though the depth of analysis is another matter. 

30 years ago, the Internet became available to a lot of people. About 20 years ago, mobile phones started to be adopted quickly. I believe AI will have a much shorter timeframe for adoption than what we saw with smartphones or the Internet. The quickest areas [of adoption] will likely be the automation of administrative work and content creation for teachers. So what used to be done manually by teachers will be done by AI. 

Analyzing massive amounts of new data points will allow educators to do personalized coaching with individualized lesson plans. This means concepts such as real time textbooks, or just-in-time learning plans will become the new normal. I see schools that are already getting rid of the paper-based textbooks and are switching to just-in-time or more dynamic real-time textbooks. 

Additionally, in the future, the student assessment will happen at a much more personalized level and in a more sophisticated way as student learning tasks will become more complex. We will do away with trying to memorize things or answering generic questions. 

Do you think about AI and learning differently across different end markets of K12, higher education, corporate learning and lifelong learning, and if so how?

In my view, the lower grades are still about building a general foundation for future learning. Whereas higher education and beyond is for more specialized knowledge and skill development. AI in lower grades will focus a lot on engaging in more discovery, or learning opportunities for knowledge [building]. In comparison, AI for later stages will focus a lot on creativity and innovation in unknown territories. A key priority in that segment is bringing about a higher level of efficiency in time and cost savings.