Mentoring

While scrolling insta the other night, I came across the story of Derek Li, a Chinese entrepreneur who removed his sons from traditional schools and is experimenting with AI-led education (says a lot about my social media algorithm). These kinds of posts and inquiries make me angry (and fearful of Skynet). I immediately started thinking, “Can algorithms truly replace the human teacher? Can education become more efficient, or even more effective, by replacing teachers with algorithms?”

To be clear, there are many areas where AI can help make education more efficient (and maybe more effective): creating learning progressions based on the standards, lesson planning, scheduling, brainstorming ideas, communications, possibly even tracking student progress (without using student names or other identifying information). But the idea of replacing teachers with AI, indeed anxiety inducing, is misguided and oversimplified. Because, as most educators know, at the foundational level of all effective learning are relationships, connections, emotional and social development, and human judgment… all the things no algorithm can reliably replicate.

What AI Can Improve

When it comes to education, AI can play a strong supporting role in the following areas:

• Lesson planning & content generation: AI can assist in structuring units, generating ideas, adapting content to different levels.
• Administrative and scheduling tasks: Freeing teachers from logistical burdens so they can focus more on teaching.
• Feedback loops and assessment data: AI tools can help identify gaps, track progress, offer drill-practice or adaptive learning. 
• Curriculum Writing, Alignment, and Improvement: In recent workshops, I’ve seen presenters use AI to identify essential standards, generate learning progressions based on those standards, and then eventually develop guaranteed and viable curricula. 

These are valuable efficiencies. But, more than anything, they are support functions. In no way are they a substitute for what good teachers do or bring to the classroom on a daily basis. 

Why Teachers Matter

Replacing teachers and the human element of school with AI overlooks very important dimensions of good education:

1. Social-emotional development & relationships
   • Before students can engage deeply with academic material, many other needs must be met (safety, connection, belonging, emotional regulation, etc.). Some call this “Maslow before Bloom,” or “relationships before rigor.”
   • Teachers see more than content delivery. They often recognize when a child is dealing with stress, trauma, anxiety, or other outside-school pressures. These sociocultural factors (family situation, socio-economic status, identity, culture) massively influence brain development and subsequent learning.
2. Mentorship, modeling, moral/character development
   • Teachers are mentors, role models, means by which students learn not just facts, but how to think, to work with others, to manage failure, to persist.
   • Human presence allows for empathy, adapting to nonverbal signals, trust (features that algorithms will always struggle to replicate).
   • I’ve seen a similar model where the ed tech entrepreneur uses “guides” to help facilitate some of the “connection” within a normal classroom. I have lots of questions about this and would be interested in seeing the data on its effectiveness. 
3. Collaborative learning, peer interaction, group dynamics
   • Lots of human learning happens in social, group settings. Students learn from one another; discussion, debate, and working in teams foster deeper understanding. If students are spending most of their time interacting with machines/screens, those peer and human interactions may be limited, reducing the benefit of collaborative learning.
   • John Hattie’s meta-analyses show that collaborative and cooperative learning have moderate effect sizes (collaborative learning = .40/cooperative learning = .55). VISIBLE LEARNING+2VISIBLE LEARNING+2
4. Long-term, holistic outcomes
   • Many qualities we deem necessary for functioning as a citizen in society, including social responsibility, creativity, resilience, and kindness, are nurtured through human relationships, interactions, and shared experiences. Algorithms can’t do that. 
   • Also, there’s still limited evidence on long-term effects of full AI-led schooling: what about emotional well-being, social skills, creativity, ethics, fairness, bias, etc. — all areas where human teachers are crucial both for guidance and oversight.

Risks & Downsides of Too Much Technology and Overreliance on AI

Most parents can probably attest to ways technology has changed their child/children. To that point, new research shows some startling effects technology (and social media) can have on children: 

• Mental health, attention, screen time, social comparison
  • In The Anxious Generation, Jonathan Haidt argues that the rise of smartphones and social media and a lack of play and human interaction with others correlates with increases in anxiety, depression, self-harm among teens, especially since around 2010. He links these changes to what he calls the “phone-based childhood,” which displaces in-person interactions, free play, and unsupervised activity. Wikipedia+2New York University+2
  • Other risks include sleep loss (and its subsequent effects), attention fragmentation, addiction to the stimulus of screens, social comparison, and loneliness. Unplugged Canada+2New York University+2
• Reduced opportunity for embodied, real-world learning
  • Learning that involves physical presence will always be hard to replicate with AI, including interacting with other students, hands-on work, field projects, and labs.
  • Play, social games, imaginative play, risky play, and collaborative projects are foundational, especially for younger children. 
  • Also, Hattie’s list considers “technology with high school students” to have a smaller effect size than human relational or instructional strategies. VISIBLE LEARNING+1
• Bias, equity, fairness, cultural context
  • AI systems are designed by people and trained on data; they can and do embed biases. Currently, they’re not very good at picking up cultural or individual differences, or responding sensitively to them.
  • Also, would this reliance on AI/algorithms widen equity gaps because not all students have equal access to technology or robust internet/devices? 
• Screen fatigue & disengagement
  • Students may be less motivated if too much is mediated by screens. Authentic human feedback, encouragement, peer interaction often drive motivation more than automated systems.

Conclusion

AI is powerful. It’s smart. And there are things it can do to make many parts of education more efficient. But efficiency alone is not enough. If we lose what makes education human, including relationships, empathy, belonging, and social learning, then we risk undermining learning, not enhancing it.

While we find ways to integrate AI into our daily practices, we must always remember that there’s no substitute for great, human, in-person teaching. Because in the end, good education is not just about what students learn, but also about who they become. Teachers, not algorithms, are the way we get there.