Framing Computational Agency

Exercising economic agency fully requires navigating or controlling computational systems. Just having access to them is insufficient. We need pliable software systems that people can adapt to their ends.

We also need applications that scaffold a user’s ability, guiding them toward deeper computational confidence. Agency begins when users can bring their full selves to a system and be guided toward greater capacity.

Pliable systems offer agency when users are ready. Scaffolding systems help them get ready. Securing systems act as ongoing companions, sustaining agency by protecting users from harm and preserving empowerment in computational environments.

Interventional systems help the structurally marginalized navigate and influence domains outside of computing per se. As an example, a worker-centered platform that helps gig drivers analyze and respond to ride-sharing algorithms gives them visibility into, and influence over, opaque algorithmic management systems that determine their experience and earnings.

Finally, compositional systems help construct shared computational infrastructure to achieve communal goals. Here, while a workplace tool that enhances employee voice by surfacing collective sentiment is valuable, one that allows them to organize around those sentiments is better.

Across these system types, computational agency prioritizes the assignment of interpretive and directional power to computer users without dictating a specific outcome.