Vivek Ramakrishnan, MBA ’23: How AI Could Help Solve the School Literacy Crisis

As an educator, Vivek Ramakrishnan is familiar with the shocking statistics on literacy gaps in the United States: Eight out of 10 children who cannot read proficiently by third grade will never catch up. However, he has also seen the problems firsthand at the school he co-founded in Madison, Wisconsin, when he was 19. “The statistics are eye-popping — but when you see what that really means in person, it’s devastating,” he says.

Helping change this is Ramakrishnan’s mission. Driven by a desire to create a fairer distribution of opportunity for young people, he believes the most effective way to do this is by closing early literacy gaps. While it is known that one-to-one tutoring is the most effective way to help children learn, it is often unaffordable for the families and children who need it most. However, in scaling up the impact of tutoring, Ramakrishnan sees recent advances in artificial intelligence as game-changing.

Through their venture, Project Read, he and his cofounder John Danner — both of whom have been teachers — are turning to generative AI to design a new kind of AI reading tutor. The tutor will combine the conversational style of a human with personalized stories in decodable text (which uses phonics, words, and concepts students already understand) that are tailored to the precise level and previous reading mistakes of each child.

For Ramakrishnan, however, designing an app that is able to spark children’s excitement and imagination is critical. “The focus now is on building something kids love using,” he explains. “Because in ed-tech, even if you have the best instructional intervention, if kids aren’t engaged with it, it won’t matter.”

The Problem

In working to close the literacy gap, Ramakrishnan is taking on a daunting challenge: Some 65% of students in the US — including 83% of low-income students — cannot read proficiently. In fact, recent data from the National Assessment of Educational Progress reveals that only 33% of all fourth graders and only 19% of low-income fourth graders can do so.

If not addressed, these students are significantly less likely than others to graduate from high school or to build successful careers, says Ramakrishnan, which means that tackling the problem early is critical.

By the time children are in third or fourth grade, he explains, most of the curriculum requires baseline proficiency in reading to access the content in social studies and science and all the other disciplines. “So if a child cannot read proficiently by this point,” he says, “the learning gaps widen in both reading and other subjects.”

For students who do not read at grade level, falling behind on their studies is only part of the problem. “If you’re lacking fundamental academic skills, you start acting out because, for example, you don’t want to read out loud in class,” he says. “You start doing all these other things to cope, and it just compounds the problem.”

Since the school literacy crisis is most acute in low-income communities, the first challenge is finding a solution that can reach sufficient scale at an affordable price. The second is ensuring external tutors are aligned with classroom instruction and the progress of different students. “You have a cost problem and a logistics problem,” says Ramakrishnan.

Ramakrishnan believes Project Read can help solve both. “I want to move the needle towards the world we all want for our kids, where the zip code or family you’re born into is not necessarily indicative of what your opportunities or life outcomes will be,” he says. “I want to decouple opportunity from circumstance.”

The Solution

While the cofounders had initially included parents among their target customers for the Project Read app, they recently decided to fully focus on teachers and schools in order to ensure Project Read reaches students of all economic backgrounds. Now, they are building Project Read as an “AI co-teacher” for the Science of Reading to support elementary school teachers in Science of Reading-aligned curricula.

Specifically, Project Read’s AI co-teacher generates “decodable” stories exactly at the child’s level, based on the specific phonics skill they are learning. As the child reads the generated decodable story out loud, the AI co-teacher listens and provides feedback and encouragement in the moment to the child. It tracks their progress and their errors and generates a next story on the fly based on the child’s previous mistakes to provide more practice and feedback for the child to master that specific skill.

Many schools have recently pivoted to Science of Reading-aligned curricula, which place a heavier emphasis on explicitly teaching phonics and decoding. The successful implementation of such curricula places a higher load on teachers, who have enthusiastically welcomed the AI co-teacher help in their classrooms.

Ramakrishnan believes the key to Project Read’s success will be in the level of customization that the technology enables and the ability of generative AI to create a virtual tutor that can interact conversationally with students and encourage them as they work through the content.

Ultimately, the co-founders aim for AI-powered tutors to engage with students when they reach stumbling points and encourage them to keep working through difficulty. “The prototype we’ve built does that. It can say: ‘You did an awesome job, but you missed one word. I highlighted it in yellow. Can you try that again?”

Advances in artificial intelligence have made this conversational interface possible. “That’s an inflection point relative to the natural language processing we’ve had up to now,” says Ramakrishnan.

However, Ramakrishnan knows what will ultimately make Project Read succeed both as a business and as a solution to the school literacy crisis: “You need to find something that kids and teachers love using, and that is instructionally effective,” he says.

The Innovator

If Ramakrishnan has seen the value of education in his work as a teacher, he has also seen it through the eyes of his parents, Indian immigrants who met in Wisconsin when his mother was doing a PhD in electrical engineering and his father was a professor of computer science.

While they placed a high value on education, his parents also wanted their children to follow their passions. For Ramakrishnan, that meant going into education. “I didn’t take the traditional path of sons and daughters of the Indian diaspora,” he says. “I’m a schools guy.”

Being a “schools guy” has served him well in his role as an ed-tech entrepreneur working to improve early childhood literacy. “I can see around corners that might take longer for someone who has not been in the space,” he says. “And I have networks of teachers — that’s super helpful.”

Additionally, his years working as a teacher, as well as his own experience going to a very diverse public high school, are powerful motivators as he builds Project Read.

Through both his school friends and his students, Ramakrishnan has seen how access to education and opportunity shapes outcomes later in life. “It felt like where you came from was a bigger predictor than who you were as a person,” he says. “Seeing friends sink or swim because of factors beyond their control just struck me as profoundly unfair,” he says. “That was what lit the spark and what still makes me tick.”

Ramakrishnan’s determination to help transform access to opportunity means he is committed to developing Project Read, even if his startup does not ultimately succeed. “Project Read could be a game changer for millions of students. That’s the dream,” he says. “But if we aren’t the people who crack solving the literacy gap at scale, maybe we move the needle far enough for another company or nonprofit to pick up the baton and solve it — to me, even that feels very worth doing.”