High school counselors in the U.S. now serve roughly 385 students each. In many districts, counselors spend less than a quarter of their time on college advising, leaving students with minimal one-on-one support as they think about life after graduation.
To fill that gap, some schools are turning to AI chatbots. The idea is simple: every student deserves more time and support in planning their future. With budget and staffing constraints, AI is a promising solution.
Imagine a junior named Mia. Late one night, she opens her laptop and starts typing into a counseling chatbot. She tells it she’s the first in her family to apply to college, that she’s nervous about life after high school, and that she’s confused about financial aid. The bot asks thoughtful questions: What subjects excite you? What kind of environment do you want? Are you nervous about leaving home? Mia answers honestly. The chatbot responds with empathy and offers practical next steps. Mia feels seen. She has access to more information than she’s ever had before. Fifteen minutes later, she has a list of colleges that seem like a good fit and a clear plan for what to do next.
This is real progress. Mia got the guidance she desperately needed, at the moment she needed it, without having to wait weeks for an appointment. She has information at her fingertips to make informed decisions about her future. That matters.
Mia leaves the conversation with answers. She also leaves alone.
The Trap We’re Walking Into
In our first episode of the Learning Curve conversation series, Brian Lightfoot said something that reframes everything: “It’s not just what you know. It’s not just who you know. Really, it’s who knows you.”
In our second episode, Julia Freeland Fisher, Director of Education at the Clayton Christensen Institute, built on this idea: “Relationships are the original social capital.” Relationships open doors to opportunity. Students need relationships with people who see them through the seasons of their lives. Networks that sustain them when they hit barriers no chatbot can anticipate.
Julia posed a critical question during our conversation: “If a bot does that instead of a human, is that one less person invested in your future?”
If a bot handles Mia’s college questions instead of a human counselor, that’s one less person who knows Mia’s name. One less person who might remember her when a scholarship opportunity comes up six months later. One less person invested in her future.
AI can generate thoughtful questions and empathetic responses, but it can’t introduce you to someone who might change your life. That’s the quiet danger of this moment: we’re building tools that replicate connection without actually creating any.
Here’s the equity trap: AI might provide low-income students with better information than an overstretched system can offer, but if we give them bots while wealthier students still have access to human networks, we’re not leveling the playing field. Affluent students will always have access to people in their target industries. If we resolve the guidance crisis with chatbots, we’re leaving the most crucial ingredient, human investment, unequally distributed.
What Students Actually Need
The stakes are high: by 2036, 70% of jobs in Texas will require post-secondary credentials. Only 27% of Houston high school graduates earn those credentials within six years. Students need access to information about their options, and AI can deliver that at scale in ways our current system cannot.
Social capital, however, matters just as much as access to information. Having a network of people willing to advocate for and connect you to others is one of the most significant predictors of economic mobility. With AI we are building systems that scale information while quietly eroding the human connections that turn that information into actual opportunities.
Students need what researchers call webs of support: peers, near-peers, and adults, each offering different kinds of help and guidance. Right now, schools treat relationships as inputs, things that help you achieve a score, instead of outcomes worth building systems around. When relationships are just inputs to academic success, we don’t plan for them to last beyond intervention. We don’t train teachers to stay in touch. We don’t reward principals who ensure students graduate with a network. We don’t create systems that allow relationships to outlast the school year.
Economic mobility unfolds over years, not semesters. The moment students really need those connections is when they’re looking for internships and jobs, long after they’ve left our classrooms. Without intentional systems to build and maintain those webs, students graduate alone.
What This Could Mean for Houston
Houston has organizations that connect students to mentors, careers, internships, and professional networks. They serve as bridges to experiences schools cannot always generate alone. The challenge is coordination. When community partners and schools work together intentionally, opportunity expands.
This requires systems that support and reward relationship-building. Educators need time and structures to connect students to these organizations and maintain those relationships beyond a single school year. Campus leaders need resources to create multi-year advisories and alumni networks. System leaders need to treat social capital as an outcome, measure whether students stay connected years after graduation, and evaluate AI tools by asking whether they serve human connection or replace it.
The Bottom Line
AI in schools is neither good nor bad. If we deploy chatbots to provide every student with access to better information while simultaneously investing in human networks, we could move toward achieving equity. If we use AI as a substitute for the hard work of building relationships, we’ll widen the gaps we’re trying to close.
The measure of equity is both how much information every student has access to AND how many people are willing to act on their behalf.
