Historically, AI has done well only when not near humans. AI for Society (AI4S) group at the AI Institute is focused on enabling people to make rational decisions despite real-world complexities of poor data, changing goals, and limited resources by augmenting their cognitive limitations with technology. Lead by Prof. Biplav Srivastava, the group works in neuro-symbolic methods, trusted AI, and applications of data-driven methods to society. We innovate as well as apply AI techniques for hard problems facing the society with an inclusive, value-driven, focus.
Our differentiating initiatives are in working with, including training, large language models and creating ontology for automated planning (generalized planning), black-box assessment and rating for AI (GAICO and ARC tools), group recommendation, metacognition-based SOFAI tool to combine fast (e.g., LLMs) and slow thinking (e.g., reasoners), and exploring AI like chatbots in trust-sensitive domains, viz. elections, opinion networks, and traffic (for SC).
Two papers accepted at 36th International Conference on Automated Planning and Scheduling (ICAPS) on learning transition models for generalized planning and solving composition of AI trust ratings as a probabilistic planning problem.
Two papers accepted at AAAI 2026 Spring Symposium on Machine Learning and Knowledge Engineering for Knowledge-Grounded Semantic Agents (MAKE) on neurosybmolic, safe, chatbots and ontology for multi-agent planning.
A book titled 'Assessing, Explaining, and Rating AI Systems for Trust - With Applications in Finance', is being released by Springer Nature ( Amazon).
Kausik Lakkaraju defended his dissertation, "Rating AI Models for Robustness Through a Causal Lens", on Feb 04.
A demo paper accepted at the ACM Web Conference 2026 on AAAI 2026 on ARC: A Tool to Rate AI Models.
Our group will be attending AAAI-26 to present the following works:
-
GAICo: A Deployed and Extensible Framework for Evaluating Diverse and Multimodal Generative AI
Outputs
(IAAI/AAAI-26)
Jan 23, Friday -
GAICo: Demonstrating a Unified Framework for Multi-Modal GenAI Evaluation
(Demo)
Jan 23, Friday - Booth 12 (Hall 3) -
OMEGA: An Ontology-Driven Tool for Explaining Multi-Agent Path Finding
(Demo)
Jan 24, Saturday - Booth 6 (Hall 2) -
Chatsparent: An Interactive System for Detecting and Mitigating Cognitive Fatigue in LLMs
(Demo)
Jan 23, Friday - Booth 5 (Hall 2)
