Explainable and Safe Chatbot in AAAI 2023 and UMBC SURFF Award

Powerful AI systems should be developed only once we are confident that their effects will be positive and their risks will be manageable. from “Letter on Pause Giant AI”

This trust must be well-founded because guaranteeing system safety calls for more than just increasing precision, efficacy, and scalability. It also calls for making sure that systems are resilient to extreme occurrences and keeping an eye out for unusual and risky behavior.

• ALLEVIATE came up in AAAI 2023 Demo
• Awarded UMBC seed faculty award for developing a Safe and Explainable Conversational Agent with focus on question generation.
• Amazing work from my student Surjodeep Sarkar in putting together the first critical review of AI-based conversational agents, conversational datasets, and evaluation metrics, from the perspective of safety and explainability.