According to Sal Khan, whether AI leads to a dystopian or utopian future is up to us.
“Technology in all of human history [has] always amplified human intent. AI is no different. Are we going to say, ‘let’s also use it to amplify positive intent?’” Khan said to a packed Shanahan auditorium. “In many ways, the education lane in AI is the most poetic one, because we have a shot of using artificial intelligence to improve human purpose and human intelligence.”
On Sept. 10, Sal Khan, the founder of Khan Academy, presented the first lecture in the annual Nelson Distinguished Speaker Series at Harvey Mudd College: “AI Won’t Destroy Education, It’ll Save It.” He discussed Khan Academy’s expansion to the realm of AI through a personalized tutoring tool known as Khanmigo. Earlier this year, Khan published the book “Brave New Worlds: How AI Will Revolutionize Education (and Why That’s a Good Thing).”
This year’s speaker series focuses on AI and learning. Kyle Thompson, the director of learning programs at Harvey Mudd College, chose this subject as a continuation of last year’s Nelson Series theme, which focused on humanity in the age of AI.
“[The previous theme] dealt more with existential questions of what is a human person as different from an artificial intelligence entity … that’s a grand philosophical topic,” Thompson said. “So we found, what if we focused on the specific question of AI and learning? Because that conversation shows up a lot on faculty and student minds.”
Khan shared insights about how AI can most support the education system without detracting from the vital role that teachers play. By assisting with lesson planning, grading and similar tasks, teachers can more effectively use their time to connect with students.
Khan used his founding story to demonstrate that the dichotomy of technology and human elements is a false choice. Khan Academy began in 2008 when he started making educational videos for his cousins.
“My cousins infamously told me they liked me better on YouTube than in person… The human element only grew when I was able to leverage technology,” Khan said. “If I had to pick [between] ‘amazing teacher, no technology’ or ‘amazing technology, no teacher,’ I would pick ‘amazing teacher, no technology’ every time. But I think we’re in a world where we can have the best of both worlds.”
He then described Khan Academy’s new AI-powered teaching assistant, Khanmigo, which launched in May. He explained that the learning tool is programmed to respond to questions in the style of a thoughtful tutor. If a student asks for the answer, it will refuse and redirect the students toward the correct approach. Remarkably, it allows users to converse with simulations of historical and literary figures and revise essays. For professors, Khan noted the value of AI in crafting lesson plans and prompts.
“I think it’s going to be a massive connectivity tool. AI will be able to provide more transparent and consistent grading, or even hold us humans accountable as we grade things on our own biases,” Khan said.
The interface is anthropomorphized, responding with appropriate emojis and having an avatar with eyes. In fact, Khan prioritized building a personality into the tool, allowing it to engage with students’ interests in the same manner of a tutor. He acknowledged the risks of forming a connection with something that mimics humans, but argued that we can maintain a balance of AI interaction and genuine human interaction.
“If I had to pick [between] ‘amazing teacher, no technology’ or ‘amazing technology, no teacher,’ I would pick ‘amazing teacher, no technology’ every time. But I think we’re in a world where we can have the best of both worlds.”
“Learning is hard…. if it can make that person more excited about learning, if they can feel connected to the learning, I’m all for it,” Khan said. “A lot of the billion dollar companies that are putting out AIs, I don’t feel like they’re putting that energy to give any personality…If you’re careful, it has so much opportunity.”
Attendee Sarah Nathani HM ’28 said she appreciated the overall positive outlook towards AI presented in the lecture.
“It was interesting to get a view that was not as cynical as I expected it to be. Now we have the inevitability [of] AI, and what are we going to do with that?” Nathani said. “I think it provokes as many questions as it answers. It was also one of my heroes, so it was really amazing to see.”
During the “Q&A” session held after the talk, a student asked how Khan would address growing concerns about the efficiency of AI educational tools.
“70 percent of [college students] don’t place into algebra. There’s a teacher shortage right now. So I’m always on the side of ‘[what] if we could put more resources into education that are used?’” Khan said.
The final question in the “Q&A” was not a question, but an expression of admiration.
“Did you know that you’re my hero?” two students enthused. The audience cheered and applauded, acknowledging their deep respect for Khan’s work.
Harriet Nembhard, the president of Harvey Mudd College, observed that having the chance to engage in conversation about the role of technology is essential for higher education.
“I think some very powerful models [were] presented,” Nembhard said. “This forum gives us an opportunity to explore a range of ideas [and] projects into what the technologies are. It’s just the sort of thing we should be doing at Harvey Mudd.”
Facebook Comments