The Design Choice — Summary

Priya has thirty-seven students. She teaches middle school math remotely from her apartment in Baltimore. She has cerebral palsy that affects her fine motor control and her speech. After two weeks, her students hear what she is saying instead of how she is saying it. After a month, several of them have started mirroring her pacing, slowing down, leaving more space between words.

AI made Priya’s career possible. Voice-to-text handles her documentation. AI tools manage the scheduling and grading that would exhaust her fine motor capacity. The cognitive multiplier, applied to someone whose physical constraints made the old architecture inaccessible, produces something the old architecture could never have produced: a teacher whose disability is a pedagogical asset.

The design choice is this: every AI system embeds assumptions about what a normal user looks like, and those assumptions determine who the system amplifies and who it excludes. The accessibility-first design, building for the person at the margin, produces a system that works better for everyone. The disability community has known this for decades. Curb cuts, closed captions, voice control: each was designed for a specific disability and became universal because the design turned out to serve a wider population than its authors imagined.

The cognitive multiplier does not have to produce more inequality. It does so when designed for the center of the distribution and offered to the margins as an afterthought. Designed from the margins outward, it produces something different: a tool that amplifies capacities the old architecture could not see because it was not built to accommodate the bodies and minds that carried them.