Professor Metin Coşgel is piloting a new AI-integrated writing curriculum in economics, one of UConn's largest majors, with the potential to shape how writing is taught across disciplines.
Metin Coşgel, a professor of economics, is exploring how AI can enhance student learning in writing courses by bridging new technology with analog skills. (Bri Diaz/UConn Photo)
As artificial intelligence (AI) reshapes classrooms and careers alike, UConn professor of economics Metin Coşgel is asking a deceptively simple question: Can generative AI help students become better writers?
The answer, Coşgel says, lies not just in what we ask students to produce, but in how we guide them through the writing process itself.
"AI can help with writing, but students need to be able to own their work and defend it along the way, not just generate a final paper at the end because the system allows it," he says.
This fall, Coşgel will launch a redesigned version of ECON 2500W, a core writing-intensive course for UConn economics majors. Supported by a Teaching Enhancement Grant from the College of Liberal Arts and Sciences (CLAS), the new curriculum integrates AI tools with traditional instruction to help students improve their writing, understand their learning process, and graduate with the skills needed for today's workforce.
Coşgel's approach is detailed in a recent article, Teaching Economical Writing in the Age of AI: A Process-Based Framework, co-authored with economics professors Richard Langlois and Thomas Miceli. The piece outlines how AI can enhance student learning by framing writing as a structured, iterative process.
"We don't want students to be expert AI users; we want them to learn how to write well," says Coşgel. "That means connecting AI with the required analog skills that we want them to graduate with."
Teaching Writing Like an Economist
Coşgel's research interests typically have little to do with emerging technologies, focusing instead on topics like the long-term economic history of the Ottoman Empire, the relationship between state and religion, and the effects of ethnic and racial diversity on performance in America's offshore whaling industry. His path into AI began, somewhat unexpectedly, with an administrative assignment.
Two years ago, he was asked to chair a search committee for a new faculty position in economics, part of a CLAS faculty cluster hire focused on artificial intelligence and machine learning. To prepare, Coşgel began learning the basics, and soon found himself increasingly interested in how AI could be applied to teaching.
"The more I learned, the more I wanted to learn," he says.
Around the same time, Coşgel was named director of undergraduate studies in economics and tasked with rethinking how the department approaches teaching in the age of AI.
His new administrative role, combined with his growing interest in AI, sparked conversations with colleagues about how the technology could support teaching in one of UConn's largest majors. The economics program serves more than 1,000 students, all of whom must complete at least one writing-intensive (W) course.
Coşgel saw an opportunity to use AI not just as a writing tool, but to reimagine how assignments are taught and assessed in W courses.
Instead of focusing solely on a final paper, his process-based model emphasizes a series of graded steps - including outlines, drafts, discussion posts, and presentations - that give students multiple opportunities to build skills and reflect on their learning.
"We want to put students in a situation where they do their own calculations, where they understand that using AI has benefits and costs in terms of not learning skills," he says.
Coşgel tested the model in an honors course this past year, and his findings helped inform his recent article. This fall, he will pilot the AI-integrated version of ECON 2500W in two sections. By spring 2026, he hopes to roll out the model across all sections of the course.
To ensure consistency, especially in sections taught by graduate assistants, he also plans to develop a standard HuskyCT course that instructors can adapt for their classes. The goal is to give every student, regardless of section, a strong and consistent foundation in writing for economics.
Cross-Campus Collaboration
Coşgel's efforts are part of a broader campus-wide conversation about how AI is changing the way UConn teaches and learns.
This past year, he served on a CLAS task force exploring AI opportunities and concerns. He has also participated in a group hosted by the Center for Excellence in Teaching and Learning (CETL) focusing on teaching intentionally with AI, alongside faculty from across disciplines.
According to Coşgel, faculty are grappling with concerns about academic integrity, student learning, and the evolving role of generative AI in higher education. He says many people tend to fall into one of two camps: those who are highly resistant and want to ban AI from their classrooms, and those who are enthusiastic about its potential but often overlook its limitations.
"As economists, we know any time you do something there are costs and benefits, and you do something when the benefits are greater than the costs," Coşgel says. "That it makes it more important to understand the costs and benefits of AI."
Despite concerns about overreliance on AI, Coşgel believes there's reason to be hopeful. Used strategically, AI can democratize access to knowledge and help students understand complex topics more quickly, he says.
"To me, the possibilities are endless, and it's fascinating to see where it's going," he says.
