At some point in our lives, each one of us has experienced a feeling of complete awe while watching a small child expertly swipe to unlock a smart device or open mobile apps to enter their own digital universe.
Over here in the YML labs, we have long been curious of the role that emerging technologies play in learning and development for young brains.
Our curiosity started a few years back when we teamed up with education startup Montessorium to develop a full app suite aimed at empowering kids to learn at their own pace. That project turned out to be especially rewarding because we provided kids a fun way to learn everything from the alphabet to international geography. We also got some digital recognition from Steve Jobs once the app launched.
Since then, our curiosity in learning and development in early education has only grown, especially as AI development continues to advance in great strides. We decided to further explore how we could bring the power of artificial intelligence tools to education.
What research is saying:
Research shows that during the preschool years, expansive psychological growth takes place and the brain is particularly sensitive. We know that screen times can greatly affect the forming of neural pathways and the way brains develop.
But because technology changes in our society are nascent, the effects of those changes are still relatively unknown and often debated. Decades ago, researchers learned that young brains need tons of stimulation to develop normally.
As a result, parents were encouraged to expose their children to as many sensory stimulations as possible. Later, digital designs and technology became more integrated into our everyday lives. We started seeing studies suggest that children who had too much screen time were more likely to develop ADHD.
For instance, in one particular study, young mice were exposed to six hours of a light and sound show on a daily basis. Results showed that there were "dramatic changes everywhere in the brain,” Jan-Marino Ramirez, director of the Center for Integrative Brain Research at Seattle Children's Hospital, told NPR.
Results like this lead some researchers to believe that our brains being wired up all the time can’t be a good thing. We weren’t built for this kind of over-stimulation. On the other hand, some researchers believe that our brains have to evolve in the way it processes information because our world is increasingly becoming more fast-paced.
In the mice study mentioned, mice that were exposed to stimulation were able to stay calm in environments that typically stressed out those who didn’t experience as much excitement.
Leah Krubitzer, a neurobiologist at the University of California, Davis, thinks studies like this show that benefits of an overstimulated brain may outweigh its negatives. During last year’s Society for Neuroscience meeting in San Diego, Krubitzer explained that we already live in a world where overstimulation is the reality. This means our brains have, whether we like it or not, already changed. Using technology correctly, in a useful, healthy way, is just the kind of stimulation that will prepare children for an always on, fast-moving world.
Because what other option do we have? We can’t turn back the clock. We can’t teach our kids in the archaic ways our grandparents were taught. Those good ol’ days don’t exist anymore.
"Less than 300 years ago we had an industrial revolution and today we're using mobile phones and we interact on a regular basis with machines," Krubitzer said during the meeting. "So the brain must have changed."
The truth of the matter is, data on how screen time affects the brain isn’t large enough to draw sweeping conclusions. Just consider how last October, the American Academy of Pediatrics lifted its longstanding rule against any screen time for kids under two. This is a standard that had been put in place since 1999.
The current recommendation comes from the result of new research. It states that young children should get screen time to help them develop the abilities to transfer knowledge from screens to the real world.
Daniel Simmonds, a resident pediatrician at the University of Maryland in Baltimore who has a PhD in neuroscience, says the key is to stick to the middle ground somewhere between the past and the future. So let your little ones interact with technology, but don’t let AI replace social human interactions.
“So much of our brain is dedicated to sensing things and making movements around [those physical things],” said Simmonds, pointing out that the show “Sesame Street helps kids learn but it’s not going to help them learn if you just sit them in front of a TV without any human interactions.”
Further proving Simmonds’ point, a 2015 study found that when iPads were given to kindergarten students to share, those students outperformed students who had their own iPads.
The study’s researchers suspected that those results have to do with the fact that sharing an iPad boosted social interactions. This type of camaraderie is crucial for development in young children. Perhaps even more telling, students who weren’t provided iPads at all scored much lower on their end-of-year achievement test compared to students who had access to an iPad.
Breaking through archaic ways of learning
Armed with knowledge that back and forth interactions between children and a caregiver is critical to language and brain development, we built an educational app that uses machine learning and image recognition to help create engaging, interactive moments. In this particular project, when our custom built iOS app asks,"Can you show me the flag of Canada?" image recognition is then used to identify whether the child is holding up the correct index-sized flag or not. This recognition happens in real-time and is instantaneous.
Teaching children about the flags of the various countries in the world requires a bit of focus on machine learning. Achieving our goal of image classification and detection to work offline in real-time required several important decisions, like whether we wanted to go with an image classification or object detection approach. In the end, we decided object detection would allow our users the flexibility of showing multiple types of flags at once.
Other important decisions we had to make include selecting the right framework for mobile devices, the network for object detection, and feeding the right data into the network during training. All of these decisions are crucial to performance and accuracy of the end product.
Ultimately, we are excited about the beginning of our exploration and the possibilities ahead!
As humans, our history stretches back hundreds of millions of years and like all biological traits, our brains have changed just like the world around us. We can’t expect to get by on outdated ways of learning. However, the key here is to not use AI tools in a way that is meant to replace humans. The true power and purpose of technology is not to substitute for human interactions but to enhance that experience and bring to life what hasn't been imagined yet. Much like Simmonds said, the key is to take advantage of what AI development can offer, and that’s a lot when it comes to sharpening young minds through learning, interacting, and communicating.
“The integration of technology and physical learning is not new, and there’s a lot of potential for it,” said Simmonds.