Amazon has been dominating the smart speaker market. As millions of Amazon members consider buying more Alexa devices, a patent has been filed for real-time accent detection and translation.
Even though two people may speak the same language, accents can sometimes be difficult to understand. The patent describes a process for identifying the accent of a person and then adjusting the computer generated voice to produce an output with a similar accent. A system is designed to allow for a two-way conversation between users with greatly different accents, making it easier for each to understand the other by translating accents.
One of the shortcuts Amazon has taken is that location data, phone numbers, and previous correspondence are all used to help identify the correct accent to use. Only a few words would be needed coupled with the previous information to accurately determine how a user is likely to be speaking.
In theory, this could also make Amazon's smart speakers significantly easier to use and understand for those with heavy accents. Greater localization and recognition of colloquial expressions may eventually be improved as a result. Since it is just a patent filing, there is no indication that Amazon actually plans to implement the technology into future products.
Another area where this technology could be useful is in call centers. Stereotypes aside, anyone that has called a tech support number has likely experienced speaking to a representative that is extremely difficult to understand. An AI-assisted accent translator could make it possible for significantly clearer communication.
Where Amazon will end up making use of its intellectual property is anyone's guess, but there are certainly many scenarios where hearing voice in your own accent would be helpful.