In January, students in David Park’s intersession course left campus and explored how Baltimore manages electronic waste. What they saw changed how they thought about technology and civic work.
Park, a Visiting Fellow at the SNF Agora Institute, designed Using AI to Streamline E-Waste Management in Baltimore to expose students to how technology tools like artificial intelligence can inform policy planning and execution.
“E-waste in Baltimore felt right because this is the digital generation,” Park said. “They don’t remember a time without the internet or social media. I asked them, How many of you cycle through your phones every year or every other year? How many of you have a drawer of old electronics at home? Every single one of them said yes.”
Students visited the Quarantine Road Landfill and the WIN waste incinerator. At the landfill, they stood on layers of waste built up over decades. The site showed how small, repeated choices accumulate. At the incinerator, they saw what happens when items meant for recycling enter the trash stream.
Students spoke with staff from the Baltimore City Department of Public Works. Workers described their daily routines, safety risks, and the limits imposed by equipment, staffing, and time. Ideas that had seemed workable in the classroom changed once students saw how those constraints play out in the real world.
Those conversations helped shape the solutions to e-waste proposed by the students.
One team focused on battery risk and worker safety. After speaking with DPW staff, the team created a tool to support front-line workers in identifying high-risk items before they enter the waste stream and flagging handling concerns within existing workflows.
A second team worked from the resident side. They built a visual guide to help Baltimore residents identify electronic waste and find appropriate drop-off locations. Field visits and DPW feedback pushed the team to create a guide with clear instructions that mattered more than exhaustive definitions. The tool aimed to reduce contamination before items ever reach the landfill.
A third team approached e-waste as a material resource. They explored how AI might help estimate the value of metals inside discarded electronics. Conversations with DPW reframed the project. Instead of treating extraction as an abstract opportunity, the team designed around real constraints. Their prototype focused on decision support, helping staff determine when recovery would make sense and when it would not.
Students used AI tools through HopGPT, a Johns Hopkins–managed platform that provides secure access to multiple large language models. The tools helped teams organize information, test assumptions, and explore options.
“We’re all learning these tools at the same time,” Park said. “The concepts are still new. That makes listening, communication, and restraint even more important.”
As projects developed, students encountered uneven data and inconsistent outputs. Park required teams to present both what they built and what remained unresolved.
“These problems don’t have clean solutions,” he said. “But students can learn how to engage them responsibly.”
Park returned often to the idea that tools do not operate on their own. “Technology doesn’t sit outside institutions,” he said. “It gets absorbed into them. If you don’t understand how those systems work, then the technology just reinforces whatever is already there.” That understanding guided how students approached AI in the course. They treated it as something that enters existing systems, formed by people, rules, and constraints.
