At ASU GSV, one of the clearest reminders about AI in higher education did not come from a product demo, a platform announcement, or a technical roadmap.
It came from students.
During a student-centered session hosted as part of Carnegie’s AI Collaborative, young people spoke candidly about how they are using AI today. Not someday. Not theoretically. Today.
They are using it to study, write, prepare for exams, and make sense of coursework. They are also using it to process relationships, navigate conflict, think through career decisions, seek validation, practice difficult conversations, and understand themselves in a world marked by uncertainty.
That reality should shift how higher education leaders think about AI.
The question is no longer simply, “How do we prevent misuse?” or “What policy should we put in place?” Those questions matter. But they are only part of the work.
The deeper question is this: How do institutions help students build agency, judgment, connection, and trust in an AI-shaped world?
Students Are Not Using AI in One Simple Way
One of the most important takeaways from the session was that students are not a monolith.
Some students described AI as an academic tool. Others described it as a thought partner, a coach, a writing assistant, a source of emotional support, or a way to prepare for uncomfortable human interactions.
One student explained that their AI use evolved from solving homework problems to writing essays, planning college applications, navigating roommate conflict, and seeking personal advice. Another described stepping back from AI after realizing it was “over-personalizing” her and pulling her away from her own intuition.
Institutions often talk about “student AI use” as though it is one behavior that can be managed with one policy. But students are using AI across academic, personal, social, and developmental contexts. The same tool that helps one student understand a lecture may help another rehearse a hard conversation, avoid an uncomfortable interaction, or seek emotional validation.
This makes institutional AI strategy much more than a classroom technology question.
It is a student experience question.
The Promise Is Real
The students were clear: AI can be meaningfully useful.
For students who lack access to mentors, coaches, professional networks, or personalized guidance, AI can lower barriers. It can help them draft an email, understand a concept, prepare for a conversation, organize their thinking, or imagine a path forward.
One student from Iowa described how AI can give geographically isolated students access to forms of coaching, support, and strategic thinking that may otherwise be more readily available to students with more privileged networks.
In academic settings, students also described a future in which AI handles more of the basic content delivery, allowing classroom time to become more relational, applied, discussion-based, and Socratic.
In that vision, faculty are not replaced. Their role becomes even more important. They question. They mentor. They challenge. They build relationships. They help students develop judgment in ways technology cannot replicate.
That opens the door to a powerful possibility for higher education.
AI could help students access support more quickly, learn in more personalized ways, and build confidence before entering moments of real human interaction.
But only if institutions are intentional.
The Risks are Developmental, Not Only Academic
The risks students named were not only about plagiarism or academic shortcuts. They were about human development.
Students talked about AI’s ability to affirm them too easily. To remove too much friction. To help them avoid difficult conversations. To become a first stop for emotional processing before friends, family, faculty, or mentors. To create a relationship that feels safer, easier, and more immediate than the human relationships around them.
One student framed the concern clearly: if AI creates frictionless environments where people are constantly affirmed, what happens to the emotional muscle required for difficult conversations, disagreement, accountability, and repair?
That question should matter deeply to colleges and universities.
Higher education has always been about more than content acquisition. It is also about becoming. Becoming more thoughtful. More capable. More connected. More resilient. More prepared to participate in communities, workplaces, and civic life.
If AI removes every point of friction, students may lose opportunities to build the very capacities higher education is meant to develop.
But if institutions ignore AI, ban it without conversation, or treat it only as a compliance issue, they risk losing students’ trust. They also miss an opportunity to shape healthier and more purposeful patterns of use.
Policy Must Be Co-Created
Near the end of the session, students were asked what advice they would give to university presidents, educators, and system leaders.
Their answer was unsurprisingly not “write stricter rules”.
It was to listen. To create spaces for honest conversation. To stop making assumptions about how students are using AI. To involve students in the policy-making process. To build structures that help students understand their own boundaries, values, and agency around AI.
One student pointed to the importance of rethinking orientation and rewriting honor codes with students, rather than only for students. The point was not that every institution needs to take the same action. The point was that AI is moving too quickly for static, top-down policies to be sufficient.
Institutions need living conversations.
They need mechanisms for students, faculty, staff, and leaders to revisit what is working, what is not, where trust is breaking down, and where new guidance is needed.
That work is difficult. It takes time. It requires humility. It may require institutions to revisit policies they only recently created.
But students made a compelling case that the work is necessary.
Human Connection Is an Intelligence
Perhaps the strongest idea from the session was this: human connection is an intelligence.
That idea reframes the AI conversation for higher education.
If AI can support writing, summarizing, coding, planning, tutoring, and analysis, then the distinct value of higher education may depend even more on the human capacities students build while they are there.
Connection. Judgment. Dialogue. Empathy. Discernment. Conflict resolution. Belonging. The ability to sit with uncertainty. The ability to learn from other people, rather than only from personalized outputs.
One student challenged institutions to help students understand what it feels like to be fully seen, heard, valued, and connected, and to treat that form of connection as something worth cultivating with the same seriousness as academic or career preparation.
That is a meaningful leadership challenge.
It asks colleges and universities to think differently about what they measure, what they reward, and what they design for. Not only average exam performance or attendance, but the quality of relationships students build. Not only whether students can use AI productively, but also whether they can remain grounded in their own values, communities, and human relationships while doing so.
The Leadership Opportunity
AI is not waiting for higher education to finish its committee work.
Students are already experimenting. Faculty are already adapting. Staff are already encountering new questions. Policies are already being tested by real behavior.
The institutions that move forward well will not be the ones that claim to have all the answers. They will be the ones that create better ways to ask the right questions.
Questions Higher Education Leaders Should Be Asking About AI
- How are students actually using AI?
- Where is it helping them build confidence, access, and agency?
- Where is it weakening trust, connection, or developmental growth?
- What should remain human?
- What kinds of friction are essential to learning?
- How should policies evolve as student behavior, faculty practice, and technology continue to change?
These are not only technology questions. They are presidential questions. Cabinet questions. Mission questions.
They require leaders to work across silos, listen closely to students, learn from peers, and build institutional capacity for responsible AI adoption.
That is the work of Carnegie’s AI Collaborative: helping higher education leaders make sense of AI’s practical, ethical, and human implications together.
Because the future of AI in higher education will not be shaped by tools alone.
It will be shaped by the choices institutions make now about trust, connection, learning, and leadership.
Questions Higher Education Leaders Should Be Asking About AI
How are students actually using AI in higher education today?
Students are using AI across academic, personal, social, and developmental contexts, including studying, writing, preparing for exams, navigating relationships, processing emotions, and practicing difficult conversations. Institutional AI strategy must account for this full range of use, not just classroom applications.
Does AI benefit students who lack access to mentors or professional networks?
Yes. AI can lower barriers for students who lack access to coaches, mentors, or personalized guidance by helping them draft communications, organize thinking, and access forms of strategic support that may otherwise be unavailable to them.
What are the developmental risks of students relying heavily on AI?
Students have identified risks beyond academic dishonesty, including AI’s tendency to affirm too easily, remove productive friction, and become a substitute for human relationships. Over-reliance may weaken the emotional and relational capacities that higher education is meant to develop.
Why should students be involved in creating AI policies at their institutions?
Students who participated in Carnegie’s AI Collaborative session advised leaders to involve students directly in policy-making rather than writing rules for them. AI is evolving too quickly for static, top-down policies to remain sufficient, and student input helps institutions build trust and more relevant guidance.
What role do faculty play as AI takes on more content delivery?
As AI handles more basic content delivery, faculty become more important as mentors, questioners, and relationship builders. Their role shifts toward the relational and Socratic work that technology cannot replicate.
What does it mean to say that human connection is an intelligence?
The idea, raised by students in Carnegie’s session, reframes AI’s role in higher education by suggesting that the capacity for connection, empathy, judgment, and dialogue is itself a form of intelligence worth cultivating with the same seriousness as academic or career preparation.
What questions should higher education leaders be asking about AI right now?
Leaders should be asking how students are actually using AI, where it is building or weakening agency and connection, what kinds of friction are essential to learning, and how policies should evolve as student behavior and technology continue to change.
