Amira uses a speech recognition library that was specifically built using students (as opposed to Alexa, Siri, and others that were built using adult speakers.) The ability to detect dialects and accents is built into Amira's speech recognition libraries.
Additionally, because Amira is driven by Artificial Intelligence (AI) using machine learning, the more students use Amira, the more the system "learns" how to detect variances in student speech including accents and dialects.
Students with speech impairments present challenges because of the variability in the types and extent of the impairment. Amira is currently being used in several schools where these speech patterns are exhibited, but we can't predict how soon there will be sufficient data to support speech recognition for all speech impairments.
As a reminder, all recordings are saved for review. This allows for rescoring of students where they may have more extreme speech issues, and that would allow for multiple listening passes by various adults within your organization!
If you have a particular example of a student you would like us to investigate, please email the student's name, school, and class to firstname.lastname@example.org.