As Ms. Randolph looked around her 5th grade classroom, her students were busy working on research papers about biomes. Some were using traditional notes and books, while others were turning to digital tools for support. One student had opened the classroom’s new AI tutor and asked it to write the entire paper. Another was using the AI tutor to brainstorm possible ideas and organize their thinking. A third student avoided the AI tutor altogether, worried that using it might count as “cheating.”
Watching these different approaches, Ms. Randolph found herself wondering: Could generative AI truly support student learning, or might it make it too easy for students to skip important parts of the writing and research process? More importantly, she thought, What kinds of guidelines and guardrails would help students use AI responsibly while still developing their own skills, ideas, and confidence as learners?
Ms. Randolph isn't alone. Educators across the country are watching the same scene play out in their own classrooms. Schools can no longer bury their heads in the sand: they must address artificial intelligence. Even if a school chooses not to officially endorse AI tools, which is increasingly difficult as many educational applications now feature built-in generative AI, administrators and educators must recognize that students use these tools for academic, entertainment, and interpersonal purposes. Ignoring this reality constitutes educational malpractice. Providing a free and fair public education requires preparing today’s students both mentally and physically for the digital world.
Over the past decade, digital wellness education has focused on screen time, active vs. passive online use, data safety, information literacy, and addictive apps. Now, generative AI introduces new challenges to wellness education. Each area of digital wellness should be updated to address students’ wellness needs in the age of AI.
Students Are More Aware Than We Think
The best way schools can start improving digital wellness is by involving their students. In the winter of 2026, as part of our decade-long digital wellness partnership with four local public middle schools, my colleagues and I at the University of Michigan surveyed 47 6th graders to gather their thoughts on AI and wellness. The questions were both qualitative and quantitative, and students were able to respond anonymously. The survey was designed to get a temperature check of their thoughts and experiences with generative AI. We initially assumed many would express enthusiasm for AI, but most were cautious about its influence:
- Only 9 percent felt AI was good for their mental health.
- Just 9 percent said they learned more when using AI.
- A mere 6 percent felt excited about a future with AI.
- 57 percent thought there should be more laws and boundaries around AI use.
If students are asking for boundaries, schools should answer by updating how they teach digital wellness, for both academic and recreational AI use.
This data shows 6th graders are well aware that using AI carries responsibility and can affect their academic and mental well-being. If students are asking for boundaries, schools should answer by updating how they teach digital wellness, for both academic and recreational AI use. Students’ concerns point to three areas schools can explicitly focus on: cognitive, social, and emotional wellness.
1. Cognitive Wellness
Shifting the Conversation from Screen Time to Cognitive Outsourcing
For years, we have focused on how much time students spend on screens as an indicator of their physical and mental health. However, instead of focusing solely on screen time, we should consider how much thinking and problem solving students are outsourcing to AI. For instance, rather than asking how long students spent online doing homework, we should ask whether they are still engaging in productive struggle, searching for evidence, making mistakes, and figuring things out for themselves. While getting quick solutions from AI may reduce screen time, it can also diminish meaningful learning. The key is to strike a balance between using AI and engaging in mental effort.
Redefining Active Hands-On Learning to Minds-On Learning
Education experts have long emphasized that active, hands-on learning is more beneficial for cognitive development and mental health than passive consumption. And this generally holds true; for example, students gain less cognitively from watching TikTok or YouTube reels than from creating their own content or researching for a presentation. However, with AI, being “active” by simply typing prompts and copying and pasting answers doesn’t always lead to meaningful learning and may even be detrimental to student mental health. Truly active learning must be “minds-on” as well as “hands-on.” The conversation needs to shift to what minds-on learning means when using AI tools. Figure 1 shows what that shift looks like in practice, contrasting minds-on choices with the moments students hand their thinking over to AI.
2. Social Learning and Agency
Co-Construction with Humans in the Loop
Learning is social, and experts have always encouraged teachers to integrate technology with social learning in mind (to avoid isolating students with devices). Social learning becomes even more important in the age of AI. Engineers and business professionals are using AI as a thought partner to solve problems and develop innovative ideas; students should be using AI in a similar way.
But we need to be careful that students know how to do this with a critical lens and with another human in the loop. Bouncing ideas off AI can be useful for gathering information quickly and for coming up with ideas the student may not have considered. However, “talking” to AI is not the same as talking to another person. A good example of this is Ying Xu’s research on preschool-age children using AI as a conversation partner (Xu et al., 2022). While the results showed AI-paired children had greater learning gains than those who did not have a partner at all, when the child was with a human partner, the human partner offered greater benefits, providing richer language and social communication experiences. In healthy AI-supported co-construction, humans remain active decision makers while AI functions as a collaborative tool for brainstorming, organizing, revising, and refining ideas. Figure 2 shows examples of what keeping a human in the loop can look like across two common classroom tasks.
3. Emotional Wellness
Continuing Emotional Check-Ins
Conversations around students’ emotions are more important than ever, as students are now using AI as a friend, relationship coach, and therapist. In the past, students knew online advice came from machines, but now, AI responses can feel like they’re from real people. It’s imperative that students know the difference between healthy and unhealthy uses of AI for emotional advice. Below are some red flags related to AI emotional wellness that schools should address with their students, along with ways to mitigate these concerns.
Red Flag: AI Replacing In-Person Support
Young people often find it difficult to approach adults or peers for emotional support, leading them to keep their feelings inside until they reach a breaking point. Increasingly, many students turn to AI for “support” with problems or concerns. While AI can offer helpful advice, it is not a substitute for human connection. Students can use AI to help articulate their feelings, but it’s essential that they share those words with a trusted adult. Whether they choose to communicate with a person by email, text, or in person, it’s important that support doesn’t end with AI. There must always be a human in the loop.
What schools can do: Schools should advise students who engage in difficult conversations with AI to also carry those conversations into real life. Teach students a simple sequence: If step one is asking AI for advice or support, step two is taking it to a trusted person who can verify its advice.
Red Flag: Thinking AI Advice Is Authoritative and That the Bot “Knows” Them
AI provides advice solely on the information the student provides, and AI’s default settings are often supportive and friendly in tone. This can make it easy for students to view AI as a “friend” and the ultimate authority on a topic. This is especially true for younger students who may still believe in imaginary figures like the Tooth Fairy, making it easier for them to see AI as a real expert “person.” It’s important to teach students that AI simply reflects the information the student provides; it does not truly know them. Teach students that they may use AI for reflection on a difficult topic or a hard conversation, but they should use people for connection, care, and follow-up.
What schools can do: To help students avoid seeing AI as the ultimate authority or a good friend, encourage them to ask AI for multiple perspectives on a topic and compare those responses with advice from a real person. Teachers can illustrate this by having students ask about a difficult social situation to an AI chatbot, a friend, a teacher, and a counselor, and compare interactions and levels of advice.
Younger students who may still believe in imaginary figures like the Tooth Fairy might see AI as a real expert “person.”
Red Flag: Emotional Dependence
Students can develop an unhealthy reliance on the validation that generative AI provides. Frequently, AI-generated advice starts with affirming phrases such as “You are thinking correctly . . .” or “You are a good person for. . . .” While these statements can feel supportive, they may inadvertently give students the impression that their views or actions are always “correct,” potentially overlooking important alternative perspectives. Relying on AI for validation can be problematic. This can actually erode a student’s self-confidence, as they are now leaning into a machine to tell them their worth rather than relying on their own everyday experiences and validation from trusted adults and peers. This can create a dangerous dependence on the machine for their mental wellness.
What schools can do: As educators, we should remind students that while using AI for occasional reflection can be beneficial, it should not become a constant means of validation. Suggest ways to get healthy validation away from screens, such as goal-setting, working toward achievement, contributing to clubs or volunteering, developing a new skill, and reflective journaling, rather than likes, streaks, or algorithmic feedback.
Red Flag: Oversharing Sensitive Personal Information
As students start to view AI as a “friend” or a private therapist, they can easily overshare their own or others’ personal information when seeking advice. The concern is that AI systems may learn from and reuse data entered by users, meaning students’ inputs are often not safe or secure with most AI tools.
What schools can do: Advise students to keep details—such as names, locations, and specific personal experiences—general when using AI tools.
Meeting Students Where They Already Are
As AI becomes an integral part of students’ academic and personal lives, schools must update digital wellness education to address both the benefits and challenges of these tools. It is essential for students to balance AI use with active, minds-on learning and maintain strong human connections for emotional support. We know from the 6th grade survey my colleagues and I conducted that students want more guardrails and structure around their AI use. Schools can start this process by developing mini-lessons and guidelines and using the examples in this article to start conversations and dialogue, as well as to provide concrete examples of how to productively use and evaluate AI.
