SerpApi.Org
Artificial intelligence is entering classrooms faster than the rules meant to govern it. A UNESCO global survey found that fewer than 10 percent of schools and universities have formal guidance on generative AI, even as adoption accelerates across education systems.
UNESCO has already warned that the challenge is not whether artificial intelligence belongs in education, but how it is governed. In its global guidance on generative AI, the organization cautions that rapid advances in AI tools are outpacing national regulatory frameworks, leaving schools exposed to risks ranging from data privacy gaps to unvalidated educational use. The report calls for a human-centered approach, urging governments and institutions to establish clear policies, ethical safeguards, and age-appropriate standards to ensure AI enhances learning rather than undermines it.
For Kirstin Stevens, an education systems designer working at the intersection of AI and child education and development, this gap signals a bigger and largely overlooked risk. The real issue is not academic misuse or efficiency gains. It is the relationship children are forming with AI.
Stevens refers to AI as a relational technology. Because it operates through language, it mirrors the way humans build identity, trust, and understanding. In her view, this distinction changes everything about how it should be introduced in education.
This dynamic is already unfolding in classrooms and homes. Children are engaging with conversational AI systems that can recall context, simulate empathy, and adapt to individual users. Stevens warns that these interactions go far beyond traditional educational tools. They shape behavior, influence decision-making, and create emotional patterns that existing safeguarding frameworks were never designed to address.
At the same time, schools are collecting unprecedented volumes of student data. Much of it is behavioral and deeply personal. Stevens has seen firsthand how quickly systems are deployed without sufficient governance. “A lot of new online schools are being built without due regard for keeping children’s data safe,” she says. “You have to design safety at the most basic architectural level. It cannot be something you think about later. It is too precious.”
Her perspective is grounded in years of building and scaling online education platforms, including systems designed for vulnerable and neurodivergent learners. Through this work, Stevens developed a focus on designing education systems where safety, ethics, and human development are embedded from the beginning. She now continues this work through The Novacene, an initiative focused on advancing human-centered AI approaches, governance frameworks, and safer technology adoption in education.
According to Stevens, AI is compounding aspects that are already struggling within the education system. “It speeds up everything that is already making things difficult for both students and educators,” she explains. “We focus on marking faster or generating lessons quicker. That gives speed to what is already under pressure.”
Policy discussions have yet to catch up. Most frameworks focus on productivity, curriculum integration, and digital skills. Stevens believes this approach is incomplete. It overlooks relational dependency, emotional influence, and the long-term developmental impact of interacting with systems that simulate human connection. Traditional safeguarding models were built for static tools and content delivery. They do not account for AI systems that learn, respond, and evolve with the user.
In response, Stevens is advancing a broader concept of literacy. AI literacy, in her view, must include relational literacy, discernment, and human judgment. Students need to understand how these systems influence their thinking, how emotional boundaries are formed, and how identity can be shaped through ongoing interaction. “We form our identities through language,” she says. “So every educator has to be part of this. It cannot sit with the IT department.”
Her work also challenges assumptions about how young people engage with AI. Contrary to popular belief, she has observed resistance among some students. “There is a real reluctance from certain children to engage with it,” she notes. “They are aware of its environmental impact. They do not want something else to speak for them. We need to listen to that.”
For Stevens, this highlights a broader failure to include children in decisions that directly affect their development.
The solution, she argues, is not to slow innovation but to redesign its foundations. Human-first AI design must become standard practice in education. That includes clear boundaries that prevent emotional dependency, robust data protection frameworks, and governance models that prioritize child development over convenience. AI should enhance human relationships rather than replace them.
Stevens is already applying these principles in the development of new learning environments. In her approach, AI is introduced first as a tool for educators, enabling them to design more meaningful and adaptive learning experiences. Student access is carefully staged and contextualized within a broader understanding of intelligence. “Education is where you discover who you are in relation to the world,” she explains. “If we outsource that too early, we risk losing something fundamental.”
The urgency of this shift is growing. Governments are still drafting policies while deployment continues at scale. Decisions made today will shape how an entire generation understands trust, knowledge, and connection. Stevens believes the moment demands a deeper level of reflection across institutions. “We are not evolving at the speed of this technology,” she says. “We need time to integrate what it means.”
Ultimately, her argument is grounded in a simple but powerful premise. AI reflects human systems and human values. Its risks and benefits are inseparable from the way it is designed and implemented. “AI is not something separate from us,” she says. “It is a reflection of us.”
For educators, policymakers, and developers, the challenge is clear. The future of education will not be defined by choosing between humans and machines, but by what can evolve in the space between. It will be shaped by whether technology strengthens human connection or erodes it.
Related Articles
