What's Blocking AI from Going Beyond Chatbots? (www.santoshkumarradha.com)

A recent exploration into Human-Agent Experience (HAX) highlights the urgent need for a shared design framework in AI interactions, akin to established theories in Human-Computer Interaction. While the tech community has developed robust design languages for interfaces across domains, HAX remains in its infancy, resulting in the premature homogenization of AI products, predominantly as chatbots. This lack of a theoretical underpinning complicates the design of diverse AI applications, from consumer assistants to complex medical systems, each requiring tailored interaction paradigms based on unique characteristics such as trust, intent clarity, and attention demands. The significance of this discourse lies in its potential to shape the future of AI design and user experience. By proposing a first-principles inquiry into the structural properties of HAX, the research seeks to identify key dimensions of interactions that vary over time—transforming user relationships with AI from static points to dynamic trajectories. Understanding these dimensions—ranging from temporal structure to autonomy levels—could lead to more adaptive interfaces that enhance user trust and efficacy in AI collaboration. As the technology evolves, establishing a comprehensive design theory for HAX could mitigate the risks of designing ineffective or untrustworthy systems and ultimately dictate how users will relate to AI in the coming years.