“How will we garden the future of our abilities to speak with computers?”

Hillary Juma spoke about how each of us has a different experience interacting with voice recognition systems. The challenges we face in our interactions with voice recognition software differ significantly, for example some accents aren’t understood by systems as well as others because they were not included in training data sets.

Being diagnosed with dyslexia prompted Hillary to interact more with the voice-to-text systems and explore various voice recognition systems. In doing so, she realized their shortcomings and has become a “cyber doula with a specialism in linguistics” to acquaint people with these systems and their complexities.