In our "Humans in the AI Loop" event series, we hosted a conversation on Governing the next Generation of AI, exploring how to bridge the gap between cutting-edge agentic AI research and the governance structures needed to ensure safety, alignment, and inclusivity.

The discussion, moderated by Mark Esposito, PhD (Chief Economist at micro1), featured Amir Afsharrad, PhD (Stanford University & micro1 Expert) and José Ramón Enríquez, PhD (Stanford HAI Researcher, Harvard University, & micro1 Expert).

Together, the speakers examined what makes agentic AI distinct and why it matters. Amir Afsharrad offered a technical perspective, drawing on his Stanford research on AI alignment to discuss jailbreak attacks, human-in-the-loop safeguards, and the boundaries of programming autonomous systems. He also touched on robotics and the challenges of governing AI systems that act in the physical world. José Ramón Enríquez brought the governance lens, sharing insights from his work at Stanford HAI on data equity, democratic oversight, and scalable civic engagement. He highlighted platforms that allow anonymous, safe participation for underrepresented voices, the importance of multilingual model training, and the use of AI tools like LLM-powered chatbots to foster reflection, learning, and inclusion.

The session underscored the need to balance technical innovation with policy frameworks, demonstrating how human-in-the-loop strategies, data norms, and democratic engagement can help shape agentic AI to serve society responsibly.