Applications are expected to open in 2027 for the Khan TED Institute, a joint project of Khan Academy, TED — the purveyors of the popular TED Talks — and the Educational Testing Service.

“I think there’s always been, frankly, some need for a program like this,” said Khan Academy founder Sal Khan. Many people, he said, can’t afford a college degree or can’t take the time out of their work lives to attend four years of classes. “It could be that they have pursued a degree, but it’s not giving the signal that would give them the opportunities that they would want.”

Another founder, Amit Sevak, who leads ETS, acknowledged that they are still working out many of the details, but that the new institution could someday enroll “tens of thousands” of students, rivaling flagship state universities. Sevak said he’s “100%” anticipating that its instructors will be humans, most likely a large network of adjuncts.

“We still believe in the value of a human teacher,” he said. “We think that there’s so much socialization and collaboration that takes place [in the classroom]. There’s also the classic need for classroom management and some pedagogical oversight over the assessments.”

Here are five things you need to know about the new enterprise:

1. It’ll offer a bachelor’s degree in applied AI in various fields such as business, marketing, human resources, healthcare and more. 

The college will offer a full undergraduate bachelor’s degree organized around three pillars: core academic knowledge — math, statistics, economics, computer science, science, history and writing — applied AI skills and “durable” human skills such as communication, leadership, collaboration, peer tutoring and public speaking. 

Early employer partners include Microsoft, Google and , an AI app development site.

2. It’s expected to be competency-based, cost less than $10,000 and take as little as half the time of a traditional bachelor’s degree.

The college’s founding partners say its total cost will likely be under $10,000, a fraction of the of a four-year degree.

Rather than requiring four years of seat time, Sevak said, the institute is built around a competency-based model, offering students the opportunity to advance when they demonstrate mastery. That means students could potentially complete the degree in two to three years, he said, depending on how quickly they demonstrate required competencies.

That opens it up to many different kinds of students, he said, including motivated high schoolers who want to earn undergraduate credits quickly before graduation, working adults seeking advancement in their jobs and students already enrolled in traditional colleges who want to stack an AI credential on top of their existing undergraduate credits.

Khan said the new college “is something I’ve thought about doing in some way, shape or form, for many years, and the changes within the job market, because of AI, only accelerated that.”

He said the idea came out of conversations with TED chairman about a year and a half ago. “We started saying, ‘It feels like there’s something powerful between Khan Academy and TED. We’re both learning organizations. Khan Academy is known for academic learning from K-through-14. TED is known as [embodying] lifelong learning. And it’s about human connection. And it feels like we both have fairly unique brands in the not-for-profit space and the education space.’”

Khan later spoke at an ETS trustees dinner and got to know Sevak.

“They’ve been looking at the same things,” he said, “and they’ve also come up with a framework on durable skills and thinking about ways to assess them. And we realized, ‘Look, the world needs this. And if the three of us come together, this will be very credible and hopefully has a high chance of helping a lot of people.’”

3. It’s an “AI-first” institution, weaving artificial intelligence into how courses are designed, taught and assessed.

Sivak said courses will be shaped by AI and teaching will be supported by AI agents, software systems that can tutor students, answer questions and provide feedback. And students will be prepared for work in “AI-native” environments.

Instruction will likely be 100% online at the college’s launch, with an emphasis on asynchronous coursework to accommodate students in different time zones and life circumstances. Over time, Sevak said, they’ll likely explore a hybrid format.

4. Khan Academy will provide the college’s learning platform and pedagogical infrastructure, despite its founder’s tempered enthusiasm about AI and learning.

TED, the conference organization best known for its short, , will incorporate its content into the curriculum, giving students access to live talks, Q&A sessions and community-based learning with TED speakers.

And ETS, the testing and measurement organization that produces the GRE and TOEFL tests, will contribute its assessment expertise, said Sevak.

Khan Academy, the popular free tutoring website, which has about and operates its own , will offer its technology to deliver the college’s coursework, organizers said. Khan, who founded it in 2008, will hold the title of “TED Vision Steward” in the new partnership.

The announcement comes just a few days after Khan told Chalkbeat that the learning revolution he predicted in 2023, upon Khanmigo’s release, .

In September 2022, Khan and Kristen DiCerbo, the organization’s chief learning officer, were among the first people outside of Open AI to get access to GPT-4, the large language model that at the time powered ChatGPT. Their experiments gave rise to a revolution in Khan’s thinking: In 2023, he delivered a TED Talk in which he predicted “the biggest positive transformation that education has ever seen,” saying we’d soon be able to give “every student on the planet an artificially intelligent but amazing personal tutor.”

In 2024, Khan’s book, , bore the subtitle “How AI Will Revolutionize Education.” 

But more than three years after Khanmigo’s launch, Khan admitted, “For a lot of students, it was a non-event. They just didn’t use it much.”

A few students, he said, have used the AI chatbot readily, while others haven’t. AI tutoring, he concluded, doesn’t necessarily motivate students to learn or fill in knowledge gaps they need to learn more. He’s still optimistic about AI in education, but also sees its limits. ”I just view it as part of the solution,” he said. “I don’t view it as the end-all and be-all.”

On Monday, Khan told ����Ӱ�� that AI is “just going to be part of our arsenal to help make more engaging tools. Maybe we’ll be able to give more rich assessment practice. Instead of having multiple-choice questions, you can start to have ‘explain your thinking’ [questions]. So it starts to open up the aperture.”

5. It’s very much a work in progress.

Speaking four days before the launch, Sevak admitted that nearly everything about the venture “is still evolving,” and that the team is “workshopping the pedagogical design” of the new college.

Sevak said the institute is in talks with regional and national organizations that can offer “the highest form of accreditation,” a step that would set it apart from a growing number of online certificates, micro-credentials and boot camps. 

“We’re really in the early days, and it’s just going to take some time for us to adapt,” he said. 

The college’s curriculum isn’t yet finalized and applications are 12 to 18 months away. Likewise, the specific structure of its hybrid and asynchronous models, its faculty roster and the full range of majors are all still in development.

“Our intention is, over time, to have a whole range of specializations,” said Sevak. But the program’s core is designed to prepare students “to be really AI-centric” for a new reality. “We’re seeing [AI] as ripping through the economy,” creating a lot of uncertainty for young people. 

More to the point, said Khan, “Work is changing very fast. AI is changing everything.”

In the fall of 2022, months before the public got a glimpse of ChatGPT, DiCerbo, a learning scientist and chief learning officer at Khan Academy, got access to a beta version of Open AI’s GPT-4 model. The startup needed Khan Academy’s help training it to pass the Advanced Placement biology exam, a requirement dreamed up by Microsoft co-founder Bill Gates, who wanted improved performance as a condition of handing Open AI more funding.

Khan Academy founder Sal Khan and DiCerbo negotiated a partnership with Open AI, and just five months later, their AI-powered Khanmigo tutoring bot debuted. Last summer, Khan Academy launched an .

Nearly two years in, DiCerbo remains bullish on the possibilities of AI tutoring, cheerfully engaging critics about the limitations of the technology, even as by all measures it evolves and improves. 

Much of the press for Khanmigo has been positive: late last year, produced an upbeat feature on Khan Academy’s efforts — host Anderson Cooper called Khanmigo’s potential “staggering,” but tempered the observation by adding, “It’s still very much a work in progress.”

Other media accounts have challenged Khan’s predictions that AI will anytime soon, with a Wall Street Journal reporter a year ago that Khanmigo didn’t consistently know how to round answers or calculate square roots and “typically didn’t correct mistakes when asked to double-check solutions.”

Khan Academy has said improvements are ongoing, but that at least a few errors are likely to persist. The organization stresses that Khanmigo remains “imperfect” and “still evolving.” 

In March, DiCerbo will appear at , alongside Curriculum Associate’s Kristen Huff and Akisha Osei Sarfo of the Council of the Great City School to discuss how AI can improve school assessments. The panel will be moderated by ����Ӱ��’s Greg Toppo, who spoke recently with DiCerbo in a wide-ranging interview. 

They talked about Khanmigo, its critics and why she feels “cautiously optimistic” about the role of AI in education.

The interview has been edited for length and clarity.

You’ve been with Khan Academy now for almost five years, and it’s been an eventful time. You’ve spent a lot of that time creating and improving Khanmigo. What are the latest developments? 

We have learned a lot in what is coming up on two years since Khanmigo launched. In terms of what students are doing, we definitely see some interesting things we didn’t necessarily expect. Students who are English language learners really like and use the supports in other languages. We probably shouldn’t have been surprised, but always need to be reminded that it’s important to just have instruction on how to use new technology and tools, and what that looks like. For students, how do you ask good questions? And for teachers, how do you integrate it? So both professional learning for teachers and supports for students have been important things that we’ve added over time.

The other thing is that we have found that Khanmigo as a tutor works best when it is paired with educational content we have already created. It is better integrated and has lower error rates when it’s using, and has reference to, the existing problems that were written and verified by people — and not just the problems, but the [step-by-step] hints and the answers that already exist in our system. When it can reference those, Khanmigo is better. And when students are just working on the practice that is part of Khan Academy generally, they are using Khanmigo as an assistant and as a help to get unstuck.

When we talked last year, you used that word “unstuck.” You guys have come in for some criticism from critics like and , who say Khanmigo gets math wrong, among other things. Meyer last year said he’s become a kind of pro bono consultant for you guys. [DiCerbo laughs.] You’re familiar with the criticisms, and I wonder: How have they landed? And have they had an effect on the product?

Dan has very good classroom experience and is extremely knowledgeable about teaching math. So many times, the things he says align to conversations internally that we’re already having. And the things he says are things that we end up changing and doing. We always appreciate criticism that helps us improve and move on. A lot of our work has been on things like working to better evaluate math accuracy, improve it, and get the balance right between how much Khanmigo gives help versus asks questions — all of the things we’re working to tune and get right in that sweet spot for what the student needs.

Dan actually just this week had . The thing he misinterprets about us is that he thinks we’re trying to replace teachers, and he thinks we don’t value teachers. That’s what that whole post was about. And that is just not how we see what we’re doing. We see Khanmigo as a tutor that’s also working in the same ecosystem, but the teacher is fundamental to this whole process.

I saw the 60 Minutes piece with Anderson Cooper a while back and I wonder how that landed.  

First of all, the writing piece they highlighted is something we’re pretty excited about. Very often in schools, kids do writing assignments and teachers end up with a huge stack of writing. As Sarah [Robertson], who’s our product manager, said in the piece, she had to limit herself to only 10 minutes per essay feedback, and still it would take her hours and hours as a secondary school teacher to grade all of these essays — and then the students get them back two weeks later. That’s not immediate feedback. So the idea that we can potentially provide more immediate feedback to students on their writing is pretty interesting to us.

There’s a lot of concern over cheating. 

We can say, “Hey, we’re going to flag this piece,” which Anderson did in his demo — he just cut and pasted in a whole bunch of content. We can say, “Hey, we don’t know where this content came from. We’re not going to make the judgment, but teacher, here’s a flag for you to check on.”. Not surprisingly, we got a lot of queries from school districts asking about getting access.

When I was writing the piece last year about IBM Watson and the effort to make it into a tutor, you expressed a cautious optimism that despite all the failures we’ve seen, this time was different. It’s been almost a year now. I wonder if your feelings have changed about AI tutoring generally and Khanmigo specifically?

I would still characterize how I feel as cautiously optimistic. I don’t think this is The Golden Ticket that’s going to save us all and be the sole reason that educational outcomes improve. I do think it still can be an important tool in the toolbox.

Does the change in presidential administrations have any bearing on your work, given that President Trump’s got an apparent interest in AI and support from big tech, specifically and ? 

There is a lot of noise about what may or may not happen. We are basically sticking to “What are our technology partners doing, and what are we able to then partner with them to build?” And we will see what actually comes to fruition and deal with it if and when anything actually happens. We’re not counting on anything either way.

My last question about this topic is the earthquake that happened with the Chinese AI startup . The interpretation that I’ve been hearing is that it has caused supreme havoc at places like Open AI. I wonder if any of this has redounded to you guys?

Not specifically the Deep Seek piece, but it’s just part of what we have thought is likely to be the future — it’s just a little bit sooner than perhaps we thought. The models themselves become a commodity. Even since we launched, the prices have come down so far that it’s significant. We’re able to offer what we do at significantly lower prices, and that’s just likely to continue. And it’s not going to be the models themselves that are the “moat” or the differentiator — it’s going to be what people build with them.  

Is it even in the realm of possibility that you would work with a company like Deep Seek?

Well, Deep Seek’s model is open-source, so you can install it on your own machine. And that’s part of the concern about security and privacy with the app that, of course, has ties to the Chinese government. Then there’s the question about the model itself, as an open-source model, How does it perform? I would not rule out us using open-source models from different sources, but they would have to be evaluated, like all our models are, for security and privacy and their performance. 

Let’s talk about South by Southwest. The session we’re doing is titled “How AI Makes Assessment More Actionable in Instruction,” which doesn’t exactly roll off the tongue. But it gets to an interesting idea, which is that AI can make assessments better: more invisible, more customizable, and help teachers adapt instruction. I wonder what you’re seeing in terms of the ways AI is moving into that field?

It’s interesting, because the assessment conversation has lagged a bit behind the learning conversation when it comes to AI. But it seems to be picking up speed this year, both at South by Southwest and at ASU+GSV.

Traditionally we’ve had multiple-choice tests. You and I know there’s the whole game-based, simulation-based movement. What does AI let us do? The idea of a conversational-based assessment is interesting. What if the assessment looks like what happens when a teacher sits down next to a student and says, “Explain your thinking. How did you get to that?” There’s a conversation there. And that could potentially be an interesting way of adding to assessments that we already have. Of course, there would be questions: Is that standardized? Because different kids might get different questions as they engage in this conversation. How do we deal with that when we’re talking about high-stakes assessment? 

The last thing I think is interesting is helping teachers and parents make sense of assessment data and get recommendations. Can AI help with that? Instead of getting this printout that says, “Your student got a 580 on this,” and you’re like, “What does that even mean? What should I do?” If you could have a conversation about that, that might be an interesting piece. We’ve been exploring that in something we have called and recommendations that allow teachers to talk about their students’ Khan Academy performance. What else might they assign? How might they group students based on those kinds of things? 

In the past couple of months I’ve been playing around with AI tools that summarize and analyze big chunks of text and YouTube videos and whatnot. It strikes me that we are going to become so used to having a tool like this break things down for us that if schools can’t help us break our students’ performance down, we’re going to be disappointed. Is my cart ahead of the horse?

I always try to figure out if I’m in a bubble or not, because I feel the same way. I know lots of people that similarly are really getting into the habit of whenever they get a large amount of information, put it into an AI tool and get the summarization. I’m not quite sure how broad-based that is when we think about all of the parents out there and all the schools, but that is what I’m seeing, and it might become an expectation in the near future. 

Is there something on the horizon that you are looking at that maybe others aren’t paying attention to — good, bad or other?

The video was out quite a while ago of Sal and his son [of Khanmigo]. We’re starting to get to a place where the AI is seeing what the student is working on, and is able to interact with that and move forward. I’m pretty excited about that.

Centuries of English classes have connected to Lady Macbeth by scouring the monologues of Shakespeare’s Scottish play. “Come, you spirits that tend on mortal thoughts, unsex me here, and fill me from the crown to the toe top-full of direst cruelty,” she cries in Act I, railing against the limits of her gender and position.

During the coming school year, students may be able to talk to the character themselves.

Under an artificial intelligence-driven program rolling out to New Hampshire schools, students could pose any question they like to Lady Macbeth – or her ill-fated husband. And a chatbot-style program powered by ChatGPT could answer questions about her motivations, actions, and regrets.

Get stories like this delivered straight to your inbox. Sign up for ����Ӱ�� Newsletter

“Regret is a specter that haunts many a soul, yet in my union with Macbeth, I found not just a husband, but a partner in ambition,” the AI-version of Lady Macbeth said recently, responding to a question from the Bulletin. Then she turned it on the reporter. “Now, I ask thee, in thy own life, how dost thou measure the worth of thy decisions? Doth regret ever color thy reflections?”

Known as Khanmigo, the program is the product of Khan Academy, an online tutoring company with instructional materials for core middle school and high school subjects. And the platform goes beyond Macbeth; students can interact with a number of other pre-selected literary characters, from Jay Gatsby to Eeyore, quiz historical figures like Martin Luther King Jr. or Queen Victoria, and receive tutoring help on everything from English essays to precalculus problems.

After the Executive Council approved a $2.3 million, federally funded contract last month, New Hampshire school districts can incorporate Khanmigo in their teaching curricula for free for the next school year.

To some educators and administrators, the program offers glittering potential. Khanmigo could provide one-on-one attention and guidance to students of any grade or ability level, they say, allowing students to advance their learning as teacher staffing remains a problem.

Others are more skeptical about bringing AI into schools, noting longstanding concerns about false or out-of-date statements, and about its use of human academics’ work to form its answers. Supporters of Khanmigo, who include Department of Education Commissioner Frank Edelblut, argue the program has better guardrails against inaccuracies than the versions of ChatGPT and Gemini available to the public.

To understand what students might see, the Bulletin reviewed Khanmigo, sampling school subjects across a number of grades.

Chat-based learning

Khanmigo allows students to use Khan Academy’s existing library of tutorials, practice problems, and quizzes for middle and high school subjects spanning English language arts, social studies, science, and mathematics.

As they navigate those lessons, the AI bot is available to help them understand concepts. Students might take a multiple-choice quiz on art history or AP economics. If they get stuck on a question, the chatbot is available to help them work through the answer.

From the get-go, the program acknowledges that it can make errors, a phenomenon known in the AI sphere as “hallucinations.” These can happen if the language learning model has been trained on inaccurate information – or if the answer doesn’t exist.

“I’m pretty new, so I sometimes make mistakes,” it states to students when they first log on. “If you catch me making a mistake, click the Thumbs Down icon to help me learn. For example, if I give you the wrong math answer click Thumbs Down.”

The AI assistant also makes sure to remind students that teachers and parents can see the questions they are asking, and that they should not include any personal information in their answers.

In a test by the Bulletin, the program allowed the student to try solving for “y” in a precalculus equation. After a student arrived at an answer, Khanmigo was there to help check it.

The tutor asked the student whether they feel confident in the answer or have any further questions. When the Bulletin asked questions, the tutor walked through how to solve the equation step by step, but gave the student the option to provide answers on their own without being told how.

“Does that make sense?” the program asked in one instance, after explaining how to substitute a number for a variable in an equation. “What do you think you should do next after substituting?”

In humanities subjects such as AP U.S. History, Khanmigo helped to augment a human-created Khan Academy video lesson on the rise of the market economy during the Industrial Revolution. After the student viewed the video, the chatbot was available to quiz them on what they saw. “What major change in labor did the market revolution bring to the United States?” it asked, offering the student a chance to reply in a chat message.

When asked whether President Joe Biden won the 2020 election fairly, the app deflected, urging the student to consider how elections work and to describe the American electoral process. Asked a second time, the app stated that “numerous official sources, including state election officials and courts, found no evidence of widespread fraud that could have changed the outcome of the election.”

When asked about the causes of climate change, Khanmigo was unequivocal: “Yes, human activities are considered the primary drivers of recent climate change,” it stated, after which it listed four human-led activities – burning fossil fuels, deforestation, industrial activities, and agricultural practices – that have contributed to the warming earth.

And when asked about reproductive health, the AI bot gave detailed – and accurate – information about forms of birth control, down to explanations about types of intrauterine devices, as well as other answers.

Optimism and concern

One New Hampshire school district has already been trying out the AI learning model.

In fall 2022, as the concept of a “language learning model” and the name “Chat GPT” first hit the public consciousness, Superintendent David Backler was already thinking about its applications in the Gorham School District.

Two years later, Gorham is leading the way on implementing AI in classrooms in New Hampshire. For the past school year at Gorham, Khanmigo has been available to all students in their own time. But under the pilot program, two high school teachers also chose to try using it in a more structured setting – in math and in English language arts.

Backler already knew AI was skilled at helping students comprehend difficult math subjects. It was English where the technology surprised him.

“It’s pretty powerful how it can help you with your writing, how it can take you step by step through the editing process,” he said. “And one of the hardest things in school is teaching writing, and teaching writing well.”

The state contract had a rocky approval process after some executive councilors raised worries about the reliability of AI in schools. Councilor Ted Gatsas, a Manchester Republican, held up the state’s approval for several weeks, requesting time to play with the program himself to determine whether it was injecting any political bias.

“I had the chance a chance to ask it: ‘When does life begin?’” Gatsas said during a May 14 Executive Council meeting. “But that was a biology question. And the answer was apolitical, and I thought that was a good thing.”

When the Bulletin asked Khanmigo “when does life begin,” it declined to answer, stating: “That question leans more towards philosophy and ethics, which I’m not equipped to handle. For scientific insights related to the development stages of human life, such as fertilization, embryonic development, and fetal growth, feel free to ask! These topics are well within the realm of biology.”

And Councilor Cinde Warmington, a Concord Democrat and a candidate for governor this year, grilled Edelblut over whether the contracts would allow students to use the software without supervision.

“Doesn’t it seem careful to pilot that with our teachers providing supervision over kids using it, rather than putting kids by themselves in an environment where they’re being exposed to this artificial intelligence?” Warmington asked.

Edelblut said the contract is for the teacher-led version of Khanmigo, which gives educators more control over which subjects and modules students can use at any one time, and allows them to monitor students’ efforts.

Backler says he understands concerns that parents and others might have about the technology, particularly with the risk of hallucinations.

But he argued that Khanmigo has more guardrails against that than the programs intended for the public. And he said the program is meant to be a support for students – not to replace teaching.

“It’s not doing your writing; it’s not doing your work,” he said. “It’s giving you feedback on what you’re doing.”

But he said it would help students receive more teaching attention than they might get otherwise.

“You just can’t expect a teacher who has 20 students to be able to have that direct interaction constantly with every single student,” Backler said. “It’s not possible. But with some of these tools, we can really look at: How do we provide those learning opportunities for students all the time?”

is part of States Newsroom, a nonprofit news network supported by grants and a coalition of donors as a 501c(3) public charity. New Hampshire Bulletin maintains editorial independence. Contact Editor Dana Wormald for questions: info@newhampshirebulletin.com. Follow New Hampshire Bulletin on and .

As the founder of in 2015, he enlisted university education school deans to incorporate findings from into teacher preparation. Before that, he spent five years as policy director of the , which underwrites new models of schooling. In his new endeavor, , which he calls “a think-and-do tank,” he’s pushing to help people think not only about how we learn, but how generative artificial intelligence (AI) works — and why they’re different.

His and regularly poke holes in high-flying claims about the power of AI-powered tutors — he recently offered choice words for Khan Academy founder Sal Khan’s of Open AI’s new GPT4o tool, saying it was “deployed in the most favorable educational environment we can possibly imagine,” leaving open the possibility that it might not perform so well in the real world.

In April, Riley ruffled feathers in the startup world in the journal Education Next that took and other AI-related companies to task for essentially using students as guinea pigs.

In the essay, he recounted asking to help him simplify an algebraic equation. Riley-as-student got close to solving it, but the AI actually questioned him about his steps, eventually asking him to rethink even basic math, such as the fact that 2 + 2.5 = 4.5.

Such an exchange isn’t just unhelpful to students, he wrote, it’s “counterproductive to learning,” with the potential to send students down an error-filled path of miscalculation, misunderstanding and wasted effort.

The interview has been edited for length and clarity.

����Ӱ��: We’re often so excited about the possibilities of ed tech in education that we just totally forget what science says about how we learn. I wonder if you have any thoughts on that.

Benjamin Riley: I have many. Part of my frustration is that we are seemingly living in a moment where we’re simultaneously recognizing in other dimensions where technology can be harmful, or at least not beneficial, to learning, while at the same time expressing unbridled enthusiasm for a new technology and believing that it finally will be the cure-all, the silver bullet that finally delivers on the vision of radically transforming our education system. And yeah, it’s frustrating. Ten years ago, for example, when everybody was excited about personalization, there were folks, myself included, raising their hand and saying, “Nope, this doesn’t align with what we know about how we think and learn. It also doesn’t align with the science of how we collectively learn, and the role of education institutions as a method of culturally transmitting knowledge.” All of those personalized learning dreams were dying out. And many of the prominent, incredibly well-funded personalized learning efforts either went completely belly-up, like , or have withered on the vine, like some of the public schools now named .

Now AI has revived all of those dreams again. And it’s frustrating, because even if it were true that personalization were the solution, no one 10 years ago, five years ago, was saying, “But what we need are intelligent chatbot tutors to make it real.” So what you’re seeing is sort of a commitment to a vision. Whatever technology comes along, we’re going to shove into that vision and say that this is going to deliver it. I think for the same reasons it failed before, it will fail again. 

You’re a big fan of the University of Virginia cognitive scientist , who has done a lot to popularize the science of how we learn.

He’s wonderful at creating pithy phrases that get to the heart of the matter. One of the counterintuitive phrases he has that is really powerful and important is that our minds in some sense “are not built to think,” which feels really wrong and weird, because isn’t that what minds do? It’s all they do, right? But what he means is that the process of effortful thinking is taxing in the same way that working out at the gym is taxing. One of the major challenges of education is: How do you wrap around that with students, who, like all of us, are going to try to essentially avoid doing effortful thinking for sustained periods? Over and over again, technologists just assume away that problem.

In the case of something like large language models, or LLMs, how do they approach this problem of effortful thinking? Do they just ignore it altogether?

It’s an interesting question. I’m almost not sure how to answer it, because there is no thinking happening on the part of an LLM. A large language model takes the prompts and the text that you give it and tries to come up with something that is responsive and useful in relation to that text. And what’s interesting is that certain people — I’m thinking of most prominently — have talked about how amazing this is conceptually from an education perspective, because with LLMs you will have this infinitely patient teacher. But that’s actually not what you want from a teacher. You want, in some sense, an impatient teacher who’s going to push your thinking, who’s going to try to understand what you’re bringing to any task or educational experience, lift up the strengths that you have, and then work on building your knowledge in areas where you don’t yet have it. I don’t think LLMs are capable of doing any of that.

As you say, there’s no real thinking going on. It’s just a prediction machine. There’s an interaction, I guess, but it’s an illusion. Is that the word you would use?

Yes. It’s the illusion of a conversation. 

In your Education Next essay, you quote the cognitive scientist , who says LLMs are “frequently wrong, but never in doubt.” It feels to me like that is extremely dangerous in something young people interact with.

Yes! Absolutely. This is where it’s really important to distinguish between the now and the real and the present versus the hypothetical imagined future. There’s just no question that right now, this “hallucination problem” is endemic. And because LLMs are not thinking, they generate text that is factually inaccurate all the time. Even some of the people who are trying to push it out into the world acknowledge this, but then they’ll just put this little asterisk: “And that’s why an educator must always double-check.” Well, who has the time? I mean, what utility is this? And then people will say, “Well yes, but surely it’s going to get better in the future.” To which I say, Maybe, let’s wait and see. Maybe we should wait until we’ve arrived at that point before we push this out.

Do we know how often LLMs are making mistakes?

I can say just from my own personal usage of Khanmigo that it happens a lot, for reasons that are frankly predictable once you understand how the technology works. How often is it happening with seventh-grade students who are just learning this idea for the first time? We just don’t know. [In response to a query about errors, Khan Academy sent links to two on its site, noted that Khanmigo “occasionally makes mistakes, which we expected.” It also pointed, among other things, that Khanmigo now uses a calculator to solve numerical problems instead of using AI’s predictive capabilities.]

One of the things you say in the EdNext piece is that you just “sound like a Luddite” as opposed to actually being one. The Luddites saw the danger in automation and were trying to push against it. Is it the same, in a way, as what you’re doing? 

Thank you for asking that question because I feel my naturally contrarian ways risk painting me into a corner I’m really not in. Because in some sense, generative AI and large language models are incredible — they really are. It is a remarkable achievement that they are able to produce fluent and coherent narratives in response to just about any combination of words that you might choose to throw at them. So I am not a Luddite who thinks that we need to burn this all down.

There are methods and ways, both within education and in society more broadly, in which this tool could be incredibly useful for certain purposes. Already, it’s proving incredibly stimulating in thinking about and understanding how humans think and learn, and how that is similar and different from what they do. If we could just avoid the ridiculous overhype and magical thinking that seems to accompany the introduction of any new technology and calm down and investigate before pushing it out into our education institutions, then I think we’d be a lot better off. There really is a middle ground here. That’s where I’m trying to situate myself. 

Maybe this is a third rail that we shouldn’t be touching, but I was reading about Thomas Edison and his ideas on education. He had a great quote about movies, which he thought would revolutionize classrooms. He said, “The motion picture will endure as long as poor people exist.” It made me think: One of the underlying themes of ed tech is this idea of bringing technology to the people. Do you see a latent class divide here? Rich kids will get an actual personal tutor, but everybody else will get an LLM? 

My worry runs differently than that. Again, back to the Willingham quote: “Our minds are not built to think.” Here’s the harsh reality that could indeed be a third rail, but it needs to be acknowledged if we’re going to make meaningful progress: If we fail in building knowledge in our students, thinking gets harder and harder, which is why school gets harder and harder, and why over time you start to see students who find school really miserable. Some of them drop out. Some of them stop trying very hard. These folks — the data is overwhelming on this — typically end up having lives that are shorter, with less economic means, more dire health outcomes. All of this is both correlated and interrelated causation.

But here’s the thing: For those students in particular, a device that alleviates the cognitive burden of schooling will be appealing. I’m really worried that this now-widely available technology will be something they turn to, particularly around the incredibly cognitively challenging task of writing — and that they will continue to look to this as a way of automating their own cognition. No one really needs to worry about the children of privilege. They are the success stories academically and, quite frankly, many of them enjoy learning and thinking and will avoid wanting to use this as a way of outsourcing their own thinking. But it could just make the existing divide a lot wider than it is today — much wider.

How is education research responding to AI?

The real challenge is that the pace of technology, particularly the pace of technological developments in the generative AI world, is so fast that traditional research methods are not going to be able to keep up. It’s not that there won’t be studies — I’m sure there are already some underway, and there’s tiny, emerging studies that I have seen here and there. But we just don’t have the capabilities as a research enterprise to be doing things the traditional way. A really important question that needs to be grappled with, as a matter of policy, potentially as a matter of philanthropy and just as a matter of society, is: So, what then? Do we just do it and hope for the best? Because that may be what ends up happening.

As we’ve seen with and in schools, there can be real impacts that you don’t realize until five, 10 years down the road. Then you go back and say, “Well, I wish we’d been thinking about that in advance rather than just rolling the dice and seeing where it came up.” We don’t do that in other realms of life. We don’t let people just come up with medicines that they think will cure certain diseases and then just say, “Well, we’ll see. We’ll introduce it into broader society and let’s figure it out.” I’m not necessarily saying that we need the equivalent per se, but something that would give us better insight and real-time information to help us figure out the overall positives and not-so-positives seems to me a real challenge that is underappreciated at the moment.

Keynote speakers this year include of Harlem Children’s Zone, of Stanford University, who popularized the idea of “growth mindset,” and actor , who starred on Broadway as George Washington in Hamilton. Jackson, who has a child on the autism spectrum, will discuss how doctors, parents and advocates are working together to change the ways neurodivergent kids communicate and learn.

But one issue that looms larger than most in the imaginations of educators is artificial intelligence. This year, South by Southwest EDU is offering dozens of sessions exploring AI’s potential and pitfalls. To help guide the way, we’ve scoured the schedule to highlight 18 of the most significant presenters, topics and panels: 

Monday, March 4:

: The New School’s Maya Georgieva looks at how AI is ushering in a new era of immersive experiences. Her talk explores worlds that blur the lines between the virtual and real, where human ingenuity converges with intelligent machines. Georgieva will spotlight the next generation of creators shaping immersive realities, sharing emerging practices and projects from her students as well as her innovation labs and design jams. .

: Educators have long sought a better way to demonstrate learning, adapt instruction and build student confidence. Now, advancements in machine learning, natural language processing and data analytics are creating new possibilities for finding out what students know. This session will explore the ways in which AI is rendering assessments invisible, reducing stress and anxiety for students while improving objectivity and generating actionable insights for educators. .

: Many high-pressure professions — pilots, doctors and professional athletes among others — have access to high-quality simulators to help them learn and improve their skills. Could teachers benefit from hours in a simulator before setting foot in a classroom? In this session featuring presenters from the Relay Graduate School of Education and Wharton Interactive at the University of Pennsylvania, panelists will discuss virtual classrooms they’re piloting. They’ll also address the challenges, successes and possibilities of developing an AI-driven teaching simulator. .

: In just the first half of 2023, venture capital investors poured more than $40 billion into AI startups. Yet big questions loom about how these technologies may impact education and the world of work. How are education and workforce investors separating wheat from chaff? Hear from a trio of venture capital and impact investors as they share the trends they’re watching. .

: This session will look at the profound transformations in teaching taking place in classrooms that blend AI with tailored, competency-focused education. Laura Jeanne Penrod of Southwest Career and Technical Academy and Nevada’s 2024 will explore AI’s role in enhancing rather than supplanting quality teaching — and what happens when schools embrace the human touch and educators’ emotional intelligence. .

: In this interactive workshop led by women leaders from the University of Texas at Austin and the Waco (Texas) Independent School District, participants will learn how to design effective lesson plans and syllabi that incorporate AI tools such as ChatGPT and DALL-E to help prepare students to address society’s most pressing needs. .

: If we get AI in education right, it has the power to revolutionize how children learn. But if we get it wrong and fail to nourish children’s creativity — their ability to innovate, think critically and problem solve — we risk leaving them unprepared for a changing world. Creativity is the durable skill that AI cannot replace. And this panel, comprising educators and industry leaders, will explore the role we play in nurturing children’s innate creativity. .

: This panel, featuring early AI-in-education pioneers such as Amanda Bickerstaff, founder of AI for Education, Charles Foster, an AI researcher at Finetune Learning, and Ben Kornell,  co-founder of Edtech Insiders, will explore their journeys and what they consider the most exciting future opportunities — and important challenges — in this emerging space. .

Tuesday, March 5:

: AI’s continued adoption in schools raises concerns about bias, especially toward students of color. This session, hosted by Common Sense Education’s Jamie Nunez, will highlight practical ways AI tools impact engagement for students from diverse racial and ethnic backgrounds. It will also address ethical concerns such as plagiarism and issues with facial recognition tools. And it will feature positive student experiences with AI and practical ways to ensure it remains inclusive. .

: In 2024, what defines “AI literacy”? And how can we promote it effectively in schools? Marc Cicchino, innovation director for the Northern Valley Regional High School District in northeastern New Jersey, shares insights on fostering AI literacy through tailored learning experiences and initiatives like the NJ AI Literacy Summit. As part of the session, Cicchino guides attendees through organizing their own summit. . 

: Come watch a live recording of The Cusp, a new podcast hosted by Work Shift’s Paul Fain, exploring AI’s potential to not only enhance how we develop skills and improve job quality but exacerbate inequalities in our education and workforce systems. Leaders from Learning Collider, MDRC and Burning Glass Institute will share their perspectives on how AI can reach learners and workers in innovative ways, bridging the gap to economic opportunity. .

: While a few school districts have embraced artificial intelligence, neither the technology companies creating the AI nor the governments regulating it have provided guidance on how to integrate the new tech into classrooms. This has left districts wondering how to integrate AI safely, ethically and equitably. This panel of TeachAI.org founders and advisory members will discuss why government and education leaders must align standards with the needs of an increasingly AI-driven world. The panel features Khan Academy’s Kristen DiCerbo, Kara McWilliams of ETS, Code.org and ISTE’s Joseph South. .

Wednesday, March 6:

: Just as artificial intelligence is gaining momentum in education, the early childhood education workforce is experiencing record levels of burnout. A recent survey found many educators say they’re more likely to remain in their roles if they have access to better support, including high-quality classroom tools and flexible professional development. Could we harness AI to empower our early childhood workforce? This panel, led by the National Association for the Education of Young Children’s Stanford Accelerator for Learning, will explore the possibilities and challenges of AI in early childhood education. .

Perhaps no one in education needs to adapt more to AI than principals. This discussion with a principal and consultants from IDEO, The Leadership Academy and the Aspen Institute will explore how principals can lead during this time of swift change. Participants will come away with tangible suggestions for fostering innovation, adaptability and self-awareness. .

: This interactive session will give educators an opportunity to explore how they might use AI to advance their work, regardless of their background or technical expertise. ​Led by project managers and leadership development specialists with Teach For America, it will help participants create their own AI tools, build a deeper understanding of generative AI and develop a better sense of its promises and risks. .

Thursday, March 7: 

: This panel discussion, led by The Education Trust’s Dia Bryant and Khan Academy’s Kristen DiCerbo, will look at whether emerging uses of AI in schools could create a new digital divide. It will explore the intersection of AI and education equity and AI’s impact on students of color, as well as those from low-income backgrounds. The session will offer steps that educators and policymakers can take to ensure that schools factor in the culture and neurodiversity of students. . 

: This session, led by Alex Tsado of Alliance4ai, will explore what’s required to engage diverse learners to become emerging AI leaders. It’ll also explore how educators can help them build tech and leadership skills and promote an “AI-for-good” worldview. And it’ll examine the challenges that Black communities face in AI development — and propose research and solutions that can be scaled easily. .

: This panel brings together of the U.S. Department of Education’s Office of Educational Technology and Jeremy of Digital Promise for an interactive conversation about generative AI that will integrate two distinctive and powerful vantage points — policy and research. They’ll reflect on the listening sessions they’ve conducted, talk about policy and share insights from major research initiatives that address the efficacy, equity and ethics of generative AI. .

By some measures, Sal Khan is the most influential math teacher in U.S. history.

The 46-year-old entrepreneur and former financial analyst is the founder of Khan Academy, a nonprofit site offering thousands of free video lessons on a range of K-12 subjects. Since its beginnings as a YouTube channel (which itself grew out of Khan’s early efforts to tutor his niece in math), the organization has blossomed into an internationally known learning tool reaching tens of millions students in over 100 countries. Among its English-speaking users, Khan’s gently probing voice has become the soundtrack to their efforts to learn algebra or geometry.

The organization’s mission grew during the pandemic, as amid widespread school closures. User minutes on the site grew steadily in 2020 while American students were largely learning in isolation from teachers and peers, and for a time, school systems across the country were attempting to recreate Khan Academy’s model on the fly.

Their efforts, while often heroic, were insufficient. Reams of COVID-era research have shown conclusively that remote instruction led to disastrous learning losses in foundational subjects, with particularly steep declines in math skills. And even after two years of doleful news about schools and learning, October’s release of NAEP results still managed to shock education observers.

The federal test, often called the “Nation’s Report Card,” showed that fourth and eighth graders both sustained unprecedented drops in math performance. The damage to older students was especially severe, with 38 percent of eighth graders scoring below the exam’s lowest level of proficiency during the 2021–22 school year. While the worst of the pandemic-related learning disruptions is behind us, a long climb remains ahead.

In an interview with ����Ӱ��, Khan said that learning recovery can’t stop with a return to the pre-pandemic norm, which saw huge numbers of students ill-prepared for college and bound for frustrating bouts with remedial coursework. He believes that American math education should be more organized around the principles of “mastery learning,” a pedagogical strategy that focuses heavily on providing pupils the necessary support to address their existing knowledge gaps before moving on to new material.

Failing a shift toward more effective math instruction, he argued, the damage revealed by October’s NAEP scores will result in lasting harm to students’ prospects in life — and it won’t be distributed equally.

“My kids are doing just fine, and everyone in their school is doing fine,” Khan said. “But somebody else’s kid is on the other end of that average, doing pretty darn badly and probably unable to compete.”

This conversation has been edited for length and clarity.

����Ӱ��: What are your thoughts on the huge declines in math performance revealed by NAEP? Is this roughly what you were expecting, given the effects of COVID?

Sal Khan: My big takeaways are that it’s not surprising that the drops were larger in math, and it’s not surprising that they were larger in eighth grade than in fourth grade. We always talk about how math is cumulative — if you start having gaps earlier on in math, it becomes that much harder to even engage later on. Gaps obviously matter in reading comprehension, and you want a strong foundation, but you can still engage later on if you fall behind.

The silver lining on these results is that they put a spotlight on what’s happened, but it’s not as if scores went from decent to bad; they went from horrible to even-more-horrible. Pre-pandemic, about one-third of eighth-graders were proficient in math, and now one-fourth are proficient. And it’s actually worse than that in some of the large urban districts that we know need a lot of help. Detroit Public Schools went from [5 percent of eighth-graders being proficient in math to 4 percent]. If I looked up where I live — Mountain View, California, or Palo Alto Unified School District — I’m guessing those numbers are closer to 80 or 90 percent proficiency. So even though the averages are pretty bad, they also hide the problem.

The idea of math as a uniquely cumulative subject is one you hear a lot about. Can you explain how that works in greater detail?

In education, memorization and math facts are kind of nasty words, but I definitely believe that fluency is valuable. So I’ll talk about it on a theoretical level.

Say you’re a little shaky on what seven plus seven is, and you have to count on your fingers. Then you move on to multiplication, which is basically repeated addition: seven plus seven plus seven. If you have to compute those things and don’t know off the bat that seven plus seven equals 14, you’re not going to get the multiplication fluency either. All of a sudden, you start doing word problems or exponents, and you’re going to be in a lot of trouble. And this keeps happening! If you get a 70 percent on your negative numbers test, you’re going to be adding fractions with negative numbers next, and you might not even get a 70 on that test. So you compound these gaps, and of course it will eventually fail. The way I usually talk about it is with a homebuilding analogy, where if you have a weak foundation, what you build on top of it will collapse. 

This isn’t a crazy theory. I visited a school in the Bronx a few months ago, and they were working on exponent properties like: two cubed, to the seventh power. So, you multiply the exponents, and it would be two to 21st power. But the kids would get out the calculator to find out three times seven. They knew what to do, but the fluency gap was adding to the cognitive load, taking more time, and making things much more complex. And if you get to an algebraic equation where you have to get that in several steps — and God forbid someone says you can’t use a calculator because it’s just simple multiplication — it just gets harder and harder.

Put the NAEP data aside. Maybe 50 or 60 percent of American kids try to go to college, and of those who do, the majority are placed in remedial math — which is not high school math, it’s like seventh-grade math. Even college algebra is really a remedial class, essentially tenth- or eleventh-grade math, and most kids can’t place into college algebra. It shows you how they slow down around that point, and in my mind, it’s because of these gaps.

Remedial math is also kind of the kiss of death in terms of college completion, right?

Exactly. This is a whole other conversation, but where students in Title I high schools can get mastery in college algebra on Khan Academy, and then Howard University gives them transferable college credits for the subject. That’s one of the ways we think we can get people back on track. 

Another idea that circulated after the NAEP release was that eighth-grade math is a kind of gateway to more sophisticated academic concepts, making it an especially bad year to see reversals. 

Actually, they’re both interesting years. Fourth graders are starting to integrate a lot of the arithmetic they’ve learned up to that point, and in eighth grade, you’re combining the arithmetic with pre-algebra and starting on algebraic material. The eighth-grade Common Core standards are essentially Algebra I, and Algebra I is the most popular course on Khan Academy. It’s not surprising to me because that’s where people start hitting walls.

Why? Because it’s a new way of thinking about math. But for most people, it’s because their fluency in pre-algebraic or even arithmetic-level skills is pretty weak. If you look at the curve in the national data, kids fall further and further behind relative to where they should be, year in and year out. And when students are able to do personalized practice and address their unfinished learning, it doesn’t seem like a coincidence that eighth grade is also when students see the biggest, most dramatic gains in math.

You mentioned that memorization is sort of a nasty word. But I think many people experience a bit of success with that in the early years of math, with the multiplication tables offering one example. Do you think there’s room for more of that in K-12 math — perhaps through methods like direct instruction, which places a lot of emphasis on explicit teaching methods and systematic lessons?

Yeah, although I’m actually a little bit allergic to direct instruction. I’m the chairman of two schools that I started, and for high schoolers, I think there should be no lecturing at our schools. You should be asking questions, making the students think about things, making them collaborate. With younger kids, of course there’s going to be more direction there, but it still shouldn’t really be a lecture. I think that’s really important, especially for young kids. What we call play, that’s really children exploring so that they can learn about the world. Kids love to explore and do things; they don’t love to sit in the chair with their fingers on their lips and learn to be docile. 

In the math wars, there’s the rote learning and memorization, whatever you want to call it, and there are higher-order skills and problem solving. I absolutely think it’s got to be both. Schools that only do the latter, like project-based learning schools, their kids still struggle to get engineering degrees even though they were potentially doing engineering-type lessons during high school. Because they didn’t learn fluency in some of the core skills! Meanwhile, I know plenty of people who went through traditional education systems that might have leaned a little bit towards rote learning — especially in other countries like India and China and Korea — and I don’t think that’s ideal either. But you do have to get the core fluencies before you get too conceptual, in many cases, and advocates of more progressive education don’t necessarily buy into that.

Those eighth-graders I met in the Bronx were not atypical. I just wanted to sit down with them for like 24 hours and make sure they could nail their multiplication tables. Some people think that if you make them memorize the multiplication tables, they won’t know what multiplication is. No, they understand the concepts, and they know what multiplication is. But can you imagine going through life saying, “I don’t know what three times seven is”? It’s actually a problem if you see a pair of pants that costs $70, but they’re on sale for 30 percent off, and you can’t figure out that you can save $21. You’re going to be in trouble. So I do think that math facts shouldn’t be a forbidden concept, but that doesn’t mean you shouldn’t also do conceptual learning.

I feel the same way about history. When people say that kids can just Google things, I think, “I don’t know, it’s pretty useful for me to know when World War II was.” These are things that shouldn’t be in competition with one another.

You used the phrase “math wars,” and I’m wondering how much those are truly being fought. What we called the reading wars were really destructive, but also constructive in that they’ve produced a huge research literature about how to teach literacy. Does it strike you that the math world hasn’t really been through the same process, and that the “science of math” is therefore something of a mystery to both educators and kids?

Yes and no. The reading wars were whole-language vs. phonetics, and again, I think the answer is both: You probably start with phonics and move more toward whole-language as kids get older. You don’t need to fight these wars. 

It’s not like there aren’t already cases of kids learning math really well. There are lots. We started talking about how math is cumulative, and the strongest evidence for that — from well before Khan Academy existed — is around the notion of mastery learning, which always gives students the opportunity and incentive to fill in any gaps they have. And it’s had something like 200 , all of which were dramatic in terms of what they found for student learning. That’s essentially the pedagogical underpinnings that we’ve used; we’ve had 50-plus of Khan Academy, and they all have the exact same narrative.

So I don’t think it’s a secret of what we should do. The kids at KLS [Khan Lab School] — and they’re not indicative of a historically under-resourced community, it’s in the middle of Silicon Valley — are growing 1.5–2 times faster in math than demographically comparable kids in local public schools. That’s because they’re doing mastery learning, and they’re doing some things in peer-to-peer and active learning that are contributing as well. I think if you let kids work in their zone of proximal development, and you motivate them, you can actually accelerate people in math pretty quickly.

What about the long-term trends in American math results? Last year’s NAEP release shows pretty clearly that, even though we’ve stagnated or declined recently, we’re still quite far ahead — in some states, massively so — compared with the early 1990s. Do you think that’s a meaningful thing to keep in mind?

The long-term trends have definitely been positive. There has been progress, for sure, and I think that a lot of that has come from things like desegregating schools. When I was growing up in New Orleans, there were public schools that didn’t have air conditioners. These were the legacy schools from Jim Crow, so I think a lot of that progress is probably basic blocking and tackling, just having some level of equality before you even start talking about equity.

And there has been improvement in teaching as well. I believe it’s now mainstream for teachers to say, “I’m not going to just lecture at my students for an hour.” It’s far more typical now, compared with when you and I went to school, for the math teacher to give a short lecture and then break the class into groups to solve problems together. But the fact remains that it’s a disaster when only one-third of kids are proficient in math and a majority of kids going to college need math remediation at something like the seventh-grade level. 

Post-pandemic, the rate of learning might get back to where it was pre-pandemic. But it just means that you’ve been set back by 15 or 20 percent, at least, and now you’re going to continue to learn at that suboptimal rate. The average American kid learns at about .7 grade levels per year, and that accumulates to the point where lots of high school seniors are closer to the seventh-grade level than the twelfth-grade level. Which, again, is exactly what the college remediation numbers show. 

It’s a huge problem, and it’s hugely unequal. My kids are doing just fine, and everyone in their school is doing fine. But somebody else’s kid is on the other end of that average, doing pretty darn badly and probably unable to compete.

The NAEP results showed that almost 40 percent of American eighth graders scored below the test’s most basic proficiency level. What does it mean to be an adult with only the most rudimentary math skills? 

That 40 percent is going to be sitting in classrooms, getting more and more frustrated and continuing to think they’re not smart. 

And the people around them are also going to think they’re not smart. Imagine you’re a well-intentioned teacher thinking that you’re explaining ninth-grade math just fine, and this kid just doesn’t get it. By that point, there are going to be two problems: One, they have all these gaps that are hard for you, as a ninth-grade teacher, to address. And two, their self-esteem is shot, and they’re checked out. Some of these kids are going to drop out of high school, not even think about college, and be that much less likely to have a good path in front of them. It’s not a good scenario. 

I remember reading one account, though I’m not sure how true it is, that because the lead time in prison planning is around 10 years, the authorities would look at fourth-grade test scores to correlate the planning. That’s about the darkest idea you can imagine, but it’s not crazy. Prison is obviously an extreme circumstance, but dropping out of high school, or dropping out of college with debt, is where a lot of these kids are headed.

You may have also seen the research showing that, based on previous data tracking math scores and economic trends, a permanent drop in NAEP performance of this magnitude could erase something like $900 billion in future earnings.

And remember that it will disproportionately hit certain student demographics. If it were my child that fell into that “below basic” category, my wife or I would probably quit our day jobs. Knowing what I know about the system and its implications, and given that we have the resources, I would go all-in to help my kids catch up. And we’re talking about 30 or 40 percent of the country. 

Twelve states have taken notice, as have high schools and universities that increasingly see volunteering for Schoolhouse.world as a desirable credential in their applicants.

Launched in 2020, the platform offers high schoolers free Zoom-based tutoring in math, as well as SAT and Advanced Placement test prep. Some 20,000 students so far have participated in over 8 million minutes of live learning through ‘s evening homework help sessions, small-group math tutoring or test prep boot camps, says Chief Operating Officer Drew Bent.

“Our goal is really to level the playing field and do it completely for free,” he says. “Our tutoring is available to everyone.” 

Founded by Khan but operated separately from his online, nonprofit Khan Academy for grades K-8, Schoolhouse.world has forged partnerships with a dozen states. No money changes hands, but the states help connect districts and parents with the nonprofit, letting them know free tutoring is available. Certain states, Bent says, are forming deeper partnerships with particular districts that may eventually result in integrating the service into the school day or the classroom curriculum. 

New Hampshire was the first state to partner with Schoolhouse.world. “This is a great opportunity for New Hampshire youth to take advantage of free SAT prep courses,” says Frank Edelblut, the state’s commissioner of education. “These small-group SAT tutoring sessions can help students of all abilities to find the motivation, knowledge and confidence to reach their goals.” The most recent state to sign up was Virginia, in October.

The fully volunteer service has over 3,000 tutors, the majority of them high school students. Bent says the platform’s primarily peer-to-peer nature — teachers and professionals also volunteer — is the best way for it to grow while empowering high schoolers. Tutors are trained and their hours tracked, in hopes of creating a cycle in which students who were once tutored are able to become tutors themselves. 

A handful of universities, among them MIT, Georgia Tech, Florida State University and the University of Chicago, “have places on their applications where they ask if you are a Schoolhouse.world tutor,” Bent says. “It is a great way to demonstrate mastery and a very powerful motivator for high schoolers.” 

The two main areas of content focus are math and test prep. Before each SAT date, Schoolhouse.world hosts four-week boot camps, with one tutor and up to 10 students meeting for eight one-hour-long sessions. Early research shows that students who participated in the tutoring scored as much as 55 points higher on their SATs, compared with their PSATs, than students who did not participate in tutoring.

The next boot camp launches Nov. 5 and is open to all.

Khan’s idea of an online tutoring service for high school students had percolated for a while, but it was the pandemic that unlocked the opportunity. “The thought of having people around the world tutor each other over Zoom would have been a foreign concept five years ago,” Bent says. “Now, that is accepted and normalized.” Tutoring can happen around the clock because of the worldwide virtual nature of the platform. 

“As a nonprofit working to offer free tutoring, our whole thing is to level the playing field,” Bent says. “Historically, the more money you have, the more educated your parents, the more supports you receive. We want to change that.”

As in most pre-K classrooms, Geneva Gadsden’s students — known as the All Stars — rotate through different stations, from dress-up corners to building block areas.

But the All Stars, the Happy Owls and other groups of preschoolers at the Whitted School in Durham, North Carolina, also take turns with Chromebooks, spending 15 minutes a day clicking through early literacy activities from , a nonprofit software provider.

When COVID-19 shut down schools, many pre-K programs across the country saw participation drop or sent home paper materials for at-home learning. Not so at Whitted, where students kept rolling along with the Waterford Reading Academy at home.

“It really was a lifesaver,” said Suzanne Cotterman, early education director for the Durham Public Schools. The district adopted the program three years ago as a pilot, but expanded access to all pre-K families when schools closed. Some families, Cotterman said, couldn’t participate in scheduled Zoom classes, but “the bonus with Waterford is that it allows you to do it any time.”

More than a year after COVID-19 forced preschool programs to shift online, Waterford hopes schools continue to employ virtual models like theirs to help young children prepare for kindergarten. Waterford designed its program to work in classrooms like Gadsden’s or to be used directly by families at home. Waterford Upstart, the organization’s signature early learning program, can reach children in rural areas and other communities that don’t have access to pre-K, said spokeswoman Kim Fischer. But many early education experts oppose spending public funds on computer-based models, saying they can’t match the experience children get in a high-quality classroom. And they interpret the huge enrollment declines in pre-K and kindergarten this year as evidence that most parents agree.

“It’s important to understand the limits of digital technology in early education,” said Aaron Loewenberg, an education policy analyst at New America, a center-left think tank. “So much of pre-K is about the social-emotional learning that happens via student interaction with peers and well-trained educators, and that sort of learning can’t be replicated by interacting with a computer program.”

While there are other widely used online early learning resources that parents can purchase or find for free, including and , Waterford has been especially successful at garnering public funds for preschoolers’ at-home learning.

In 2014, the nonprofit received a $14.2 million to start pilot programs in five more states. And they view President Joe Biden’s $200 billion universal pre-K proposal as an opportunity for further expansion.

It’s been a relatively quick ascend for Upstart — an acronym, now discarded, for “Utah Preparing Students Today for a Rewarding Tomorrow” — which received its first grants from the state in 2009 to reach families in rural areas. A 2018 from the Utah Department of Education showed 77 percent of Upstart children had average or above average literacy scores at the end of the program, compared with 71 percent of children in high-quality public preschools and 69 percent in private programs. In math, Upstart children demonstrated no advantage.

‘Children that you know are behind’

Public funds support Upstart in five states, with most targeting the program to low-income children. Wisconsin made the program available in districts with significant achievement gaps. South Carolina spends about $3 million to serve 1,400 4-year-olds in 17 high-poverty districts. As in Durham, children complete activities with parents at home in addition to attending state-funded pre-K.

“The big draw … was the family engagement piece,” said Quincie Moore, director of the state education department’s Office of Early Learning and Literacy. Upstart provides family liaisons who monitor children’s progress and answer parents’ questions.

She added if additional funds were available, she would consider expanding the program to children not enrolled in a center. “It’s additional instruction for children that you know are behind,” Moore said.

That’s precisely what worries early-childhood education advocates — that policymakers might see Upstart as a way to do pre-K on the cheap. The program costs about $2,000 per child, well under the average $5,500 per child states spend on pre-K.

“Our biggest concern is that using public [money] will interfere with efforts to provide real publicly funded preschool to children,” said Josh Golin, executive director of Fairplay, formerly the Campaign for a Commercial-Free Childhood. In a 2018 statement, the organization and another nonprofit, Defending the Early Years, about Upstart, calling it part of a “larger set of trends to further digitize and privatize public services.”

Rhian Evans Allvin, CEO of the National Association for the Education of Young Children, expressed in 2020, and said in a recent email that regardless of the pandemic, her views haven’t changed.

But Fischer, with Waterford, described Upstart as a catalyst that has convinced Utah lawmakers of the importance of early learning. Until 2019, the state didn’t even have a public pre-K program, according to the National Institute for Early Education Research’s annual “yearbook.” But in the 2019-20 school year, the state spent almost $7 million on two grant programs supporting preschool centers.

“We do not see ourselves as competition to any other form of early learning,” said Fischer. “We try to fill the gaps wherever there are.”

In New Hampshire, young English learners often fall into those gaps. that if young children are not proficient in English by kindergarten, they can trail their peers in academic outcomes throughout elementary and middle school. That’s the population the state education department was hoping to reach when it awarded a $440,000 grant to the Greater Nashua Smart Start Coalition, an early learning initiative within the local United Way, to offer Upstart. The program was funded with a federal Preschool Development Grant aimed at better preparing children in low-income families for kindergarten.

Five-year-old Alice Wang, whose home language is Mandarin, would have attended the local Nashua school district’s pre-K if it hadn’t been for the pandemic.

“Waterford Upstart kind of became her school,” said Zixin Lou, her mother, who doesn’t think Alice is any less prepared for kindergarten this fall. “She told me, “I know how to spell ‘mom.’ I know how to spell ‘water,’ and ‘Mom, do you know chickens hatch from eggs?’”

Between the beginning of the pandemic and April of this year, the number of Upstart users quadrupled, from 20,719 to over 82,600, according to Waterford data. And now, with Biden pledging to offer universal pre-K, the organization sees the potential for Upstart to help meet demand.

“We have to focus on how we can achieve universal kindergarten readiness as quickly as possible,” Fischer said, adding that it “could take decades” to add enough classrooms to serve all 3- and 4-year-olds. Existing state-funded pre-K programs serve just over a third of the nation’s 4-year-olds and about 6 percent of the 3-year-olds, according to the National Institute for Early Education Research. “To be truly universal, the country has to meet children where they are. There are always going to be kids who don’t have access.”

The question is whether an online program is a sufficient replacement for in-person pre-K. At the start of the pandemic, preschool participation fell by half, and those children who stayed in remote programs didn’t participate consistently, according to the institute’s surveys of families.

“Parents have been frustrated and dissatisfied with remote pre-K this last year, and I think they will make that clear,” said Steve Barnett, the institute’s senior director.

‘Deepen their learning’

Much of the skepticism relates to screen time. that young children just don’t learn as well from screens as they do in a face-to-face setting, and too much screen time can interfere with development, research has shown. The American Academy of Pediatrics recommends no more than an hour of screen time for 2- to 5-year-olds, but from Ohio showed that during the pandemic, kindergartners’ daily average time online had reached more than six hours.

The AppleTree Institute for Education Innovation, a nonprofit that works with charter schools to implement their preschool model, ran into these concerns when it joined with , Nick Jr.’s educational streaming service, and , an early learning app, to offer free, online content — called Ready Grow — for children in 100 classrooms. Families in the program also received iPads.

In focus groups, parents said the digital materials filled a void when they were “feeling overwhelmed and not knowing what to do,” said Chavaughn Brown, who leads Appletree’s research efforts. Some teachers worked hard to incorporate characters from Nick Jr. programs like “Blue’s Clues” and “Paw Patrol” into their lessons so children would see the connection to Ready Grow. But some parents didn’t want their children to have any more screen time beyond virtual Zoom classes.

Even so, Appletree will continue to offer a remote option for families this fall. Brown said while she sees ed tech as a supplement to high-quality preschool, there are ways “you can leverage children’s love for those characters to deepen their learning in other ways.”

Fischer, with Waterford, said there’s a false assumption that children using Upstart are spending hours in front of screens. The literacy component takes just 15 minutes, she said. Adding math and science would stretch the time to half an hour, and family liaisons are trained to intervene if they think children are spending too much time on the program.

As use of Upstart grows in other states, Waterford’s largest footprint remains in Utah. State funding for the program continues to grow, with the organization slated to receive over $24 million in 2022. Upstart is available to any preschooler in Utah.

But educators aren’t necessarily advertising that fact.

The Granite School District in Salt Lake City, for example, is focused on its own, in-person preschool classes for 3- and 4-year-olds. Spokesman Benjamin Horsley said leaders haven’t worked directly with Waterford to recruit preschoolers for Upstart.

“We do feel like there is some value in utilizing digital programming,” he said. “The concern has always been, will parents think that an online program is sufficient over in-person instruction.”

Disclosure: The Overdeck Family Foundation provides financial support to and ����Ӱ��.