Trustworthy AI Agents for Report Generation and Evaluation
Date:
AI agents are increasingly used to generate and evaluate reports in healthcare, government, research, and compliance. Across domains, trust hinges on accuracy, transparency, human oversight, and dedicated evaluation pipelines rather than raw model capability. In this talk, we will discuss the rise of AI report generation, trust bottleneck, and trustworthiness challenges. Then, we discuss more details about recent representative projects as case studies (FinCritic, DSM5AgentWorkflow, SciText2Eq, Multi-AgentConsensusSeeking, WikiHowAgent) to see how the trustworthy AI agents work with various domains (e.g., finance, scientific research, and healthcare).