The Future of Child Development in the AI Era (Neugnot-Cerioli, 2024) is the first publication of everyone.Ai, an organization whose mission is to ensure responsible AI for children. As noted in the extensive literature presented on human brain development, the brain continues to develop until well into a young adult’s twenties. Thus, ensuring responsible AI for children, this report is particularly centered on birth through age 25, with a heavy emphasis on early childhood and adolescence. The report highlights that most children have at least some exposure to technology by the time they are 4 months, with children often having four or more hours of out-of-school screen time per day before they are 14. In short, the need to figure out how to engage children with technology in safe and effective ways is a real and pressing need.
This report, which includes extensive review of literature on child development and educational technology combined with interviews with 16 specialists identified as being in AI, technology product development, and child development. Note that there appear to be no educational researchers, including learning scientists, in this pool. This omission is important because it leads to an overall tone throughout the document that we are fitting AI into an existing educational system that is being conceptualized as being unchanging. This has significant implications in terms of thinking about potential steps forward and potential pitfalls.
The report presents a vast body of information about child development and the ways in which technology is already known to interact with that development. The author then projects forward to what the implications of AI are for young people given what is known about child development and what is known about previous educational technology innovations.
Overall, the author moves back and forth between benefits and drawbacks, leading the report to feel relatively well-balanced. However, it is dense, does not include much elaboration, does not include educational/learning researchers, and is clearly leading to the punchline, which are a set of Propositions that, likely, will be shaping the work of everyone.Ai going forward.
Some highlights of the report include that there are some drawbacks and some benefits for AI that can be pretty well anticipated at this time. For example, AI is likely to perpetuate the problem of children spending a lot of time on screens – and for their parents to do the same, providing less opportunity for the critical, real-world interactions children need for proper development. The paper outlines that executive function, language development, and social development are all critically situated in the real-world for young learners. On the flipside, though, AI could support adapted learning, better engagement, and democratize access to high-quality materials. It is undeniable that AI can play a critical role in supporting learners with a wide variety of divergent needs. For example, work with Autistic children using robots indicates clear benefit for developing particular social skills. Similarly, work with students who are visually impaired suggests the ability to adapt learning to better support their unique challenges.
The propositions that are offered support the argument made earlier by the author. That is “…for all the promise it holds, synergistic collaboration between educational scientists and AI developers will be crucial to ensure that AI-driven educational tools are effectively tailored” for young learners (p. 18). They include such topics as the need to protect children’s rights and privacy through regulations on AI developers and rules protecting the data that can be mined from these vulnerable users. Shifting educational metrics from “engagement” to “learning”, creating curriculum for students and teachers to learn about AI ethics and literacy, and insisting on the creation of AI that does no harm are included as propositions for improving product development.
While this report provides a wealth of knowledge to project potential, often likely, benefits and drawbacks to implementing AI, it did not address the question of how AI changes what students should learn. The only hint, in fact, that the author was considering changes in education was in the proposition that we shift away from a focus on engagement to instead focus on learning. However, the report absolutely highlighted areas that will require significant changes. First, it highlights the challenges facing humans in terms of learning how to connect to each other and interact with each other. For neurotypical children, there is no substitute for actually interacting with other humans. It helps them learn critical communication skills and how to read body language. These are both critical to living on a planet with billions of other humans. Further, neuroscience has found that motor skills, executive functioning, and emotional regulation are all tightly connected in brain anatomy. Inadequate support for these things will lead to humans unsure about how to interact in productive ways. Thus, while AI may be a powerful tool for many things, young people need to have face-to-face experiences that help them develop into healthy, well-rounded adults.
A second idea about learning that was clear throughout the report was the need for new content. Research has shown that many youth graduate from high school with significant misconceptions about how technology works and that younger learners may struggle to separate fantasy from reality, particularly as they tend to anthropomorphize the technology. Add to this the emerging privacy and ethical issues that large-language models have already created, and one can see there is a pressing need for instruction about AI, not just instruction with AI. Such instruction needs to help users understand how AI works, how to protect themselves, and how to act in ethical ways with the technology.
A third idea that emerged, but that was somewhat underdeveloped was the notion that AI can change instruction in some powerful ways when its use is intentionally designed. If not intentional, it can, instead, undermine learning. As the Human Intelligence movement grows, we will be further defining some of these intentional uses and how they benefit learners. Unlike this report, we posit that gamification for the purposes of increased engagement and monitoring of learners to better understand engagement is an inadequate way of supporting learning with AI. Instead, we should turn the learning on its head. What can AI do for humans that will enrich our experience. If we consider that humans are inherently creative, pattern-finding, critical thinkers who benefit from interacting with other humans, focusing on the ways AI can take over the teacher’s role and/or extend the classroom management arm of the teacher’s role is wholly insufficient. Instead, we should be thinking about the ways AI can support learners to be creative problem solvers. For example, by offering just-in-time support in a problem-oriented curriculum or by providing students with a historical analysis of whether their “new” idea is new or if it has been done before. If we shift the role of learners away from simply accumulating information to instead interact with the world around them and we support those interactions with AI tools, young people continue to be very grounded in where they are and who they are becoming, while no longer feeling the borders of their city or country. But, this requires a rethinking of education. It requires moving away from lock-step curriculum and standardized tests. And, it requires finding innovative ways to keep our children safe as they interact with all of these tools.
Post written by Chandra Orrill