Presenter Information

Location

https://www.kennesaw.edu/ccse/events/computing-showcase/sp26-cday-program.php

Document Type

Event

Start Date

22-4-2026 4:00 PM

Description

This project addresses the urgent need for transparent chatbot safety evaluations amid rising concerns about AI-facilitated self-harm. Using public social media datasets, we simulate two tasks: (1) detecting suicidal ideation via emotion-based risk scoring, and (2) stress-testing a support-style chatbot against 888 high-risk prompts, including euphemisms and “for a story” framing. A multi-label classifier trained on GoEmotions feeds emotion profiles into a logistic regression model to generate suicidality risk scores. These scores guide a local chatbot built with Ollama’s llama3, which analyzes user messages and steers responses toward safe, empathetic behavior. Evaluation shows ~90% of replies were safe or supportive. This framework links emotional signals to risk heuristics and enables reproducible safety testing under realistic self-harm scenarios.

Share

COinS
 
Apr 22nd, 4:00 PM

EX-04-142 Modeling Distress and Evaluating Chatbot Safety for Suicide-Related Social Media Texts

https://www.kennesaw.edu/ccse/events/computing-showcase/sp26-cday-program.php

This project addresses the urgent need for transparent chatbot safety evaluations amid rising concerns about AI-facilitated self-harm. Using public social media datasets, we simulate two tasks: (1) detecting suicidal ideation via emotion-based risk scoring, and (2) stress-testing a support-style chatbot against 888 high-risk prompts, including euphemisms and “for a story” framing. A multi-label classifier trained on GoEmotions feeds emotion profiles into a logistic regression model to generate suicidality risk scores. These scores guide a local chatbot built with Ollama’s llama3, which analyzes user messages and steers responses toward safe, empathetic behavior. Evaluation shows ~90% of replies were safe or supportive. This framework links emotional signals to risk heuristics and enables reproducible safety testing under realistic self-harm scenarios.