Designing agency into assistive robotics

Before clinical neuroimaging, I worked on assistive robotics and prosthetic systems.

The domain was different, but the design question was similar:

How much control should a machine take, and how does the person understand what it is doing?

The work

The product had to operate within physical, cognitive and environmental constraints.

A successful interaction was not simply one where the mechanism worked. The user needed to feel that the machine was supporting their intent rather than replacing it.

This required thinking about calibration, feedback, physical comfort, control, error recovery and confidence during repeated use.

[ image — Prototype photo ] Photo or render of the prosthesis. If none, keep text-only.

What I carried forward

That work shaped how I later approached AI-driven clinical products.

In both domains, a machine produces or recommends something consequential. The user still needs enough visibility to understand it, enough control to question it, enough feedback to act confidently, and a safe way to recover when it is wrong.

The technology changed. The design responsibility did not.