We are in an educational crisis. 

Children are starting school without the skills they need to learn effectively. The most important skill a child will ever learn is how to read because it is the foundation upon which all other lessons can be learned. However, only 34% of children start school with the necessary understanding. This creates a knowledge gap that grows as they progress through the years. The children become disengaged due to a lack of success, resulting in 8,000 children dropping out of high school every day. Of those who manage to graduate, only 37% can read at or above proficiency.

Initially, we were taught in the home, by our family, or via a personal tutor. If you didn’t have the means, the church might offer some classes.

When the world began to industrialize and compete in the 1800s, a new model was needed to move us out of our agricultural past and into a more mechanized society. New skills were needed and an educated citizenship was required. 

At-home schooling and small classrooms began to gradually morph from their personalized roots into a more scalable educational model and the Prussian education model was adopted. The Prussian model prescribes an age-based curricular scope and sequence to scale similar education across the population efficiently. It was an industrial area invention that used the then-modern efficiency models to scale education in the same manner as production. 

I will not delve into the more controversial sides of this approach that had to do with societal control but focus instead on the unintended consequences of this industrialized learning model.

The Prussian model is based heavily on the categorization of topics and people and has proven as efficient as possible in the same way that factory assembly lines are efficient. That is, it is great at turning out units that can conform to the assemblage mechanism. 

The problem is that people are not units; we develop at different levels and at different times. Some of the reasons for this are inherent, and some are environmental. Whereas some children may be “early bloomers” when it comes to education and some may be held back for other factors such as having parents who speak in their native tongue. We do not all begin our educational careers with the exact same skill sets for many reasons.

More modern research also suggests that we all have some combination of unique learning  styles as well, such as visual or kinesthetic, that may not be well accounted for.

In the early days, this was easily accommodated as a parent or personal tutor could tailor lessons at a student’s pace and make assessments in real time to customize lessons for each student. In pre-industrialized classrooms, the teacher-student ratio was easier to manage, and teachers were unencumbered by the pressures of testing and age/stage curricula. Further, one teacher would often teach several grades simultaneously and could, therefore, group students according to ability rather than age.

As this Prussian model was adopted, the focus was to streamline the effort for efficiency, and ages seemed the best way to group students. After all, if they were all given the same exposure to the same curriculum, wouldn’t they then be about equal? It made the most sense and this approach lifted us out of our agricultural roots and enabled economies to flourish.

However, we have been following this model for the past 100 years or so in earnest, and it is proving to have some setbacks. As I mentioned, we do not all enter school with the same experiences or advantages, and it has become apparent that we do not all learn equally.

Further, the size of the classroom and the ability of the teacher can also have a great impact on the process. Human biases and shortcomings can seep into the process. Again, we are not assembling carburetors here, all uniform and equal. We are assembling humans with myriad differences.

The best way to teach is personal, with a healthy amount of personalization. When my children were very young, my wife Kim took it upon herself to prepare them for life. This included preparations for schooling. Sure, we went to the educational stores and purchased books and toys designed to give them a head start, but we didn’t observe everyone doing the same.

Additionally, Kim was very adept at taking the materials and making them fun and relevant for each of our vastly different children. She would create songs containing the content and make up stories that featured my boys and their interests. Paramount among these were the stories of the “Rocket Boys” Two young astronauts who learned all about the universe (disguised curriculum) through their adventures. My boys excelled in this environment and Kim deserves all the credit for their highly creative and fun memories of preparing for school.

When I started developing products for LeapFrog, I always remembered this approach, and as I started to employ it, I tried to understand exactly what was going on and why it was so effective. I spoke with educators and educational experts; I sought out the best practices when it came to teaching strategies. As it turns out, this is a strategy that is commonly agreed to be the most effective. 

As a member of LeapFrog’s Advanced Content Team, we would constantly challenge ourselves to create the most effective products we could and try to unlock the amazing potential of personalized and fun learning.  We were very effective, and the critical and commercial success of the company reflected it. 

We spent countless hours creating and testing different approaches to learning with children. Because everything was so meticulously tested with children, you start to get a real feel for what is going to fly and what isn’t. But, you also start to get a very clear picture that all children are different. It isn’t a good thing or a bad thing, they just are.

When we were testing various learning mechanisms for the incredibly successful Letter Factory DVD and toy line, I started the practice of getting down on the floor with the children, one-on-one at their level, to make up stories, songs, and games that they might respond to. We found many mechanisms that made that line particularly effective, but that is a whole other column. But basically, I stole Kim’s approach to teaching. I gave each child the time, attention, and compassion I would have given my own, and something startling happened, they became immensely engaged.

From that moment on, I became convinced that this personalized approach to teaching and learning should be possible for all children, regardless of background or ability. It is a better way to teach. The problem, of course, was not that no one was interested in doing it (as I mentioned, everyone agrees this is the gold standard), but that it is impossible to scale this type of personalization. As a society, we must scale. It is what propels us.

As a team, we tried desperately to include this type of learning in all products, but we could only create the illusion of personalization. That is, age-based curriculum and faux references to personal things that may be in a child’s life. But it was unreachable to make it truly personalized because technology would not allow us to generate outside of the one-to-many model.

So, when I started playing with generative AI in early 2023, I quickly realized that one of the real potentials of this technology was to, at last, enable the personalized creation of content at scale. No longer were we imprisoned by the boundaries of scaling one directional content.

Because generative AI can create text, images, and other content on the fly, based on predefined rules, it is the key to creating entertainment on the fly. As educators, we can set limits and rules, but the AI is free to make safe alterations to story elements to tailor them for each child. Imagine every story being a “choose your own adventure” or “Book of Me”. 

That is the promise of AI, not to replace the author but to allow for a world of new possibilities!

So, one part of the problem was solved: inserting the child and their preferences into the content so that it is engaging and entertaining to them personally…not just an illusion.

But, one very important piece was still missing. How would we replicate the other half of the puzzle? How would we approximate the act of getting right down there on the floor and engaging with the student at their level? This required being there, on the floor…and paying attention. And how did I do that in those early Letter Factory sessions? 

By paying great attention to the children as they read, played, and listened. Picking up verbal and non-verbal cues. Were they engaged? frustrated?, laughing?, singing with me? etc. When they verbalized (we were after all teaching phonics), how did it sound? Were they retaining the knowledge or maybe just memorizing it? What were the looking at? Clicking? Touching? These are all the cues you look and listen for.

I knew that whatever we created using AI could scale, but it would be the same hidebound paradigm that we had been using for over 100 years. It was personalized, but it was still unidirectional. It still put every kid into a bucket based on their age or stage and blasted at them. And that is when it hit me like a bolt! If the product was capable of making the same nuanced types of assessment, we could change the entire way of teaching and learning back to what we know works best. An attentive and personalized exchange of knowledge completely personalized for each student.

So we created Leaonardo. 

Leonardo uses the multi-modal signals to analyze focus, attention, and performance to understand every child at their level. Leonardo mixes that knowledge with what it knows about the child to generate highly personalized content that will engage and delight the child.

To change the paradigm back to two-way communication, we needed to include both sides of the coin. We needed to be able to react to the child in real-time in order to personalize.

That is the missing piece of the puzzle. That is what Kim was doing with our children, not just making up super creative stories and games, but being there on their level, reacting to them holistically and personally. This is the goal of Leonardo.

Leonardo combines the expertise of an amazing teacher with a parent’s focus and understanding of each child because Leaonardo is paying attention.

It all starts with the most important skill of them all, reading. Reading is the lesson upon which all other lessons are built, so it is key to learn it as early as possible. This starts with the alphabet and phonemic awareness. What do the letters look like and what are the sounds that they make? Instead of rote memorization, Leonardo is right there with the child as they progress.

Leonardo listens and learns where the child is at curricular on a sound-by-sound level. Leonardo understands what words a child is struggling with based on how many times and for how long they look at a word before attempting to decode and then breaks down their vocalizations by phonic. It pays attention to where they touch and tracks it as a signal, because this is where the child is actively asking for help.

Only with this deep understanding can Leonardo generate the next page of text, which is designed to not only be as engaging as possible but also to hit perfectly on the individual child’s cognitive ability at that moment. If the first line was “Sam has a cat”, but the child struggles to decode, then Leonardo might pull the child aside and work on simple decoding skills, or even go all the way back to individual letter sounds based on performance. Conversely, if the child is doing well, then the complexity of the words and sentences might increase. If the next line was going to be “The cat is Pat.” Leonardo may, based on the child’s current performance, instead deliver, “This is Sam’s cat, Pat” or “This is Sam’s funny brown cat named Pat” depending on all of the knowledge garnered at that point.

This is, of course, a very simple example, but the implications of it are powerful. No longer does the child have to wade through multiple levels of curriculum she may already understand to get to the next level. The next level is there, on the next page as soon as the child is ready. This is the promise of generative AI, but it must have situational awareness to be effective.

The underlying platform can grow with the child as they progress to more complicated texts. Importantly, though, the knowledge also travels with the child across the curriculum. For instance, if a child possesses a natural proclivity for math but is encumbered by reading because of environmental reasons, what will happen when they come across word problems or science problems? Normally, efforts would need to be made to ensure the child can navigate the text appropriately. In the worst case, this logic skill could go unnoticed. This is not an issue with Leonardo because the reading level is already baked into the problem.

As they grow older still, Leonardo’s underlying technology can make higher and higher levels of text more accessible. The works of Shakespeare xx, and xx, become infinitely more accessible. Leonardo can help with workplace training to ensure comprehension and therefore adherence, foreign language learning, rehabilitation, and many more areas. Anywhere people need to read and/or learn, this approach is best.

Leonardo is a refreshing new approach to education that has its roots in the tried and true methods of the past. It is successful because it has an awareness of the student and the flexibility of generative content based on rules. Leonardo is a great tool for learning and not a replacement for parents and teachers, in fact when used in concert the child will benefit incredibly.

— Christopher D’Angelo