Blind and visually impaired students face persistent barriers to STEM education. They’re less likely to participate, less likely to be engaged, and less likely to enter STEM careers. A big part of the reason is simple: most learning tools weren’t built with them in mind.

We set out to change that. With support from the National Institute on Disability, Independent Living, and Rehabilitation Research (NIDILRR), we spent Phase I of an SBIR grant tackling the two hardest technical problems: how do you navigate a 3D game world without seeing it, and how do you make a game work with screen readers when it’s running in a web browser?

What we built

We developed two core accessibility features for Tyto Online.

Navigation assist for the 3D world. We designed audio cues and voice assist to help BVI students move through the game environment — finding objectives, interacting with objects, and exploring spaces. The system uses spatial audio and pitch changes to signal direction and proximity.

Screen reader support for game mechanics. Our game mechanics — like the Argument Builder, where students construct evidence-based claims — don’t follow standard web page structure. We prototyped approaches to make these non-standard interactions work with screen readers by implementing them natively in HTML5 rather than relying on the game engine’s limited accessibility support.

What surprised us

We co-designed these features with a team of blind and visually impaired middle and high school students, and several of the things we learned challenged our assumptions.

BVI students want to explore, not just complete tasks. We expected students to want efficient navigation — get to the objective, do the task, move on. Instead, they wanted to wander. One adult tester asked to turn off the navigation sounds so they could go investigate a sound they thought was a waterfall. BVI learners want the same sense of discovery that sighted learners get from a rich game world.

Many BVI students avoid screen readers even when they need them. Our co-designers told us they don’t like screen reader voices, find the tools slow, and aren’t familiar with features like heading navigation. Some rely on extreme magnification (400%) even though it gives them headaches. Making something “screen reader compatible” isn’t enough — the experience has to be genuinely enjoyable, using natural AI voices for character dialogue while reserving the screen reader voice for interaction controls.

Students accept poor usability because they’re used to it. Our co-designers were fine scrolling through dialogue line by line at 400% magnification. When we introduced natural AI voice acting, they were thrilled — they could listen instead of laboriously reading. Look for opportunities to improve the experience even when users aren’t complaining; they’ve been conditioned to accept limitations.

100% of testers could navigate the 3D world. Both rounds of usability testing (conducted externally by the Center for Accessible Technology) confirmed that the navigation assist system worked. Adults with gaming experience found it intuitive; younger testers needed more time with the audio cues but could use the system effectively.

What this means

The findings validated our approach and shaped our Phase II plans. The bigger insight is about what equity in game-based learning actually requires. It’s not just making content technically accessible — it’s ensuring that BVI learners get the same sense of engagement, curiosity, and joy that makes game-based learning effective for sighted learners.

Accessibility isn’t a feature we’re adding on top. It’s making sure the core experience — concrete, active, joyful — reaches everyone.

Current status

Phase I is complete. We’re planning Phase II to address the remaining barriers identified by our co-designers and the C4AT usability testers, including richer spatial audio descriptions of the environment, customizable navigation controls, and scaffolding to help students build screen reader skills alongside STEM skills.