Northeastern Nursing Students Use AI Teen to Practice Substance Use Screenings

Image: Matthew Modoono/Northeastern University

Nursing students at Northeastern University are training for one of the most difficult parts of the job — talking with patients about substance use — with the help of an AI teen named Jordan.

Since fall 2024, students enrolled in the public health community nursing course have been using AI SimBot, an audio-based simulation tool that lets them conduct mock interviews with virtual patients. The goal: prepare students to ask tough, emotionally charged questions in a low-risk environment.

💬 A Smarter Way to Practice Patient Conversations

In the past, students practiced patient communication through standardized patient simulations, where actors play the role of patients. While effective, these simulations come with high costs and scheduling challenges.

According to Tiffany Kim, associate clinical professor at Northeastern’s Bouvé College of Health Sciences, these simulations require hiring actors, securing physical space, and debriefing each student individually — all of which can be time-consuming and logistically complex.

To solve this, Kim and Northeastern graduate Yash Gopalji Pankhania created AI SimBot using large language models from OpenAI. It’s now used across Northeastern’s campuses in Boston, Burlington, Fall River, and Charlotte.

🧠 How It Works: Meet Jordan and Dr. Casey

The AI simulation is split into two parts:

• First, students talk with Jordan, an AI-powered teen designed to mimic a real patient who may be struggling with substance use.

• Then, they debrief with Dr. Casey, an AI assistant that provides personalized feedback on the conversation.

Each student interview is recorded and transcribed, allowing them to upload the conversation to Canvas, Northeastern’s learning platform, for grading and self-review.

🚦Using the CRAFFT Tool

The simulation is based on the CRAFFT screening tool — a widely used method to identify risky substance use in adolescents. CRAFFT stands for:

During the interaction, students:

• Practice administering the CRAFFT questionnaire

• Build rapport through active listening, empathy, and nonjudgmental responses

• Receive structured feedback from the AI debriefer

👩‍⚕️ Student Perspective: “It Felt Realistic”

Isabelle Iannotti, a recent Northeastern nursing graduate, used SimBot during her summer term.

“The chatbot was pretty apprehensive at first to answer questions, which was very realistic,” she says. A teen at the doctor’s office for drug use would likely be hesitant.

She said the conversation was a good test of how nursing students might navigate real-life scenarios with teen patients.

Her chat with the AI debriefer also provided helpful feedback. She had jumped around while asking the CRAFFT questions, mixing them in with other conversation. The debrief made it clear that they needed to ask the questions consecutively — something she hadn’t realized beforehand.

🛠️ Open-Source and Continuously Improving

Since its launch, AI SimBot has gone through several updates and is now open-source on GitHub. Anyone can download and adapt it.

Kim says the tool provides more opportunities for experiential learning, especially when resources like actors or simulation labs are limited.

“Because students can repeat the exercise as often as needed, it reinforces competency-based learning, where mastery comes through practice, feedback, and reflection,” she explains.

🤔Nurses, have you ever practiced a difficult patient conversation through simulation? How do you think an AI tool like SimBot could support your learning? Share your thoughts in the discussion forum below!