When: Friday, October 9, 1:00 PM
Where: Tyler 106

Abstract
As AI chatbots are increasingly being used in education contexts and digital modalities are increasingly being used in depression assessment research, it is important to understand human perspectives and bias. My research started in two main areas: exploring AI chatbot acceptability at Bryant University and comparing digital depression assessment datasets. For the former research area, we surveyed faculty, staff, and students about their AI chatbot use and perceptions at the end of 2023 and 2025. While usage increased over the two years, support of AI chatbots in education remained relatively consistent. For the latter research area, we identified observational datasets for depression assessment collected by smartphones (n=80), wearable technology (n=80), and cameras (n=63). Women comprised the majority of participants across all three modalities, with patients less frequently recruited for smartphone datasets and the elderly more frequently recruited for wearable technology datasets. Missing dataset details were prevalent, raising concerns regarding the ability to assess the bias and generalizability of models constructed on these datasets. These two research areas converge in my newer research directions, which involve agentic agents and smart rings.

Bio
ML Tlachac is an Assistant Professor of Data Science and AI Departmental Fellow at Bryant University. Tlachac earned a PhD in Data Science from Worcester Polytechnic Institute and has since published 14 journal articles. Conducting interdisciplinary research, Tlachac specializes in affective computing, digital phenotyping, human-computer interaction, and AI pedagogy. Tlachac recently received an RI-INBRE pilot grant to investigate smart ring acceptability and is committed to mentoring undergraduate research assistants. Alongside research, Tlachac currently teaches undergraduate data mining and graduate natural language processing.