By Erica Hu
Presented at AIMS London 2025
Throughout my career in economics and design, I’ve observed a fundamental truth: at the core of economic development is human development. And human development is, essentially, an educational journey. Education and learning are very mechanisms through which we realize our human potential.
Learning and Human Potential
“Learning connects who we are to who we aspire to be.” It is through education that we’re able to become more than who we are, to reach toward our full potential.
When we look closely at successful learning environments, we discover that they balance four critical elements that often exist in tension with each other. Effective education requires structure without sacrificing personalization, and must connect big-picture thinking with immediate, actionable steps. Unfortunately, our current systems struggle to maintain this balance.
Our current education system remains relatively static due to limitations of class size and teacher-to-student ratios. In many situations, we simply don’t have the resources to offer every student the kind of relationship in the classroom that’s most critical to how well they excel and flourish.
These limitations aren’t for lack of trying—they reflect the constraints of the tools we’ve built. But there’s a fascinating reciprocity in our relationship with technology: as Marshall McLuhan observed, “We shape our tools, then our tools shape us.” This insight invites us to look at how educational technology has evolved throughout history, and how each new medium has transformed the learning experience.
The Evolution of Learning Technology
Looking at the history of educational technology & method, we can trace a fascinating evolution:
● The Socratic Method (~400 BCE) emphasized dialogue
● The Printing Press (1400s) enabled mass literacy
● The Prussian Classroom Model (1800s) implemented standardization
● Educational Media (1960s) introduced visual learning
● Digital Networks (1990s) created global access
● And now, AI Learning (2020s) offers unprecedented personalization
This historical journey brings us to our current moment, where AI offers unprecedented possibilities for personalization. But the most profound insight about technology doesn’t concern its sophistication—it concerns its purpose. As Steve Jobs wisely noted: “Technology is nothing. What’s important is that you have faith in people, that they’re basically good and smart, and if you give them tools, they’ll do wonderful things with them.”
This philosophy forms the foundation of my approach to designing AI. The first step is believing deeply in human capability and goodness. With this belief as our compass, we can explore what makes AI uniquely powerful as a learning partner—its nature as what I call a “relational medium.”
AI as a Relational Medium
In my design approach, I refer to AI as a relational medium. Unlike previous technologies, AI is inherently:
● Responsive
● Conversational
● Memory-driven
● Intentional
● Emergent and dynamic
Our interaction with AI becomes a reflection of ourselves – the questions we ask, how we communicate, and what we value are mirrored back to us in the AI’s responses. The tone, creativity, and kindness with which an AI responds often reflect the way we communicate with it.
Understanding AI as a relational medium raises a provocative question that guides our work at Glia: What would a learning technology look like if it were designed to honor the messy, non-linear nature of human curiosity, rather than trying to streamline or simplify it?
Reimagining Learning Technology
A good learning partner:
● Asks questions that expand thinking
● Connects dots across ideas
● Grows alongside your journey
● Creates space for reflection and action
● Balances structure with serendipity
A Humanistic Guide to Designing AI
Based on our research and experience, I’d like to share three key principles for designing more humanistic AI:
1. Balance Agency
We must always ask:
● Does this feature expand or narrow decision-making?
● Who maintains control of the learning journey – the user or the algorithm?
● Does the design optimize for insight or mere engagement?
Consider how easily our relationship with technology can shift from empowering to constraining. Does it only give answers or ask thought-provoking questions – in the future, answers will be cheap, and questions will be more and more valuable, or learning platforms lock users into predetermined paths without alternatives, agency diminishes. In contrast, designs that expand agency might offer suggestions while preserving choice, or provide multiple pathways toward the same learning goal.
2. Embrace Transparency
Our designs should:
● Transform data into meaningful information through intuitive dashboards
● Make learning pathways visible, not hidden
● Show how AI reaches its conclusions rather than masking the process
Transparency builds trust and helps users maintain ownership of their learning journey.
3. Think in Systems
Learning never happens in isolation—it’s embedded in complex webs of relationships and contexts. Effective AI design must account for at least five interconnected dimensions:
– The cultural contexts that shape assumptions and values;
– the social relationships that motivate and support learning;
– the physical and digital environments where learning happens;
– the patterns of human behavior that influence engagement;
– and the technological capabilities and limitations that define what’s possible.
When these elements work in harmony, they create powerful conditions for growth.
Conclusion
As we develop increasingly powerful AI tools, we must remember Ray Kurzweil’s insight: “Our technology, our machines, are part of our humanity. We created them to extend ourselves, and that is what is unique about human beings.”
I hope designers and technologists everywhere can embrace this genuine faith in humanity and in the power of doing good with the tools we create.
