Agent-Oriented Modeling for the Age of AI: Nine Pivots Toward a Reconceptualization of Requirements Engineering

Modeling techniques have been instrumental for conceptualizing and architecting complex systems. Since the early days of requirements engineering (RE), concepts such as process and object have served as core abstractions undergirding many modeling languages and have seen extensive application in all kinds of domains. Today, software systems are engaging with humans in ever deeper, more intricate, sometimes even intimate ways. They leverage pervasive digitalization to learn from data, acquiring increasingly humanlike characteristics. An abstract concept of agent, proposed much earlier in AI for characterizing artificial agents, suitably extended and generalized, could potentially serve as a core abstraction for requirements modeling for today’s complex sociotechnical environment. In this chapter, we outline a modeling ontology by elaborating on the sociotechnical actor concept originally proposed in i*. The proposal also suggests shifting the focus of RE from the specification of a target technology system to one of assisting stakeholders to explore how they might augment their “selves.” A personal wellness app is used as a motivating example.