One of the biggest criticisms of voice-recognition software since it first appeared on the public market has been its inability to recognize strong, local, regional dialects and accents. It seems to be a criticism that developers have taken notice of though, with many investing in regional voice talent to provide a more personalized service.
A 2017 article by WIRED highlights a lot of the problems people were having when it came to applications such as Apple’s Siri and Amazon’s Alexa recognizing their accents. The problem, they point out, is that artificial intelligence (AI) is created in such a way that it can recognize only what it’s been taught.
No wonder then that confusion can arise when it comes to making requests in strong accents. How long will it be until AI can flawlessly understand an accent from Boston, Scotland or even areas with lesser spoken dialects than that? Sooner than you may think thanks to regional voice talent…
The role of regional voice talent in artificial intelligence
Google Home is fast becoming one of the most popular smart speakers on the market. At the start of January 2018, the company boasted that it had managed to sell a Google Home device every single second since 19 October. An impressive feat which, if accurate and continues to happen, means well over 10 million units will have been sold.
It’s Google Home’s commitment to using professional voice talent that has them leading the way in the next generation of speech recognition software. A video that went viral at the end of January shows some of the impressive new features that Google has introduced to Home (including the aforementioned ability for its AI to learn from its owners).
In the video, Home is asked a question by people with two distinctly different accents and uses its machine learning capabilities to recognize the dialect and provide answers respectively. An American and a Briton ask it to spell ‘color’, with Home giving answers based on the dialect it hears and its understanding of the region. In this case either including or excluding the “u” in color.
Of course, anything Google can do Amazon wants to do better. Not only does Amazon’s Alexa in Australia speak with a native accent, but it also says regional phrases such as “You should probably chuck a sickie”.
Pair your AI with regional voice talent
It raises a smile but it’s also an incredible technological development for Amazon. They don’t want a blanket voice to interact with their consumers. Amazon want to provide the choice of communicating with AI through languages that people understand. Not only to increase personalization but also to build a trusting relationship and create a unique selling point that will elevate them above competitors in what is becoming an increasingly competitive market.
For this strategy to be effective they can’t rely on somebody doing their best impression. Companies like Amazon who are developing voice recognition are working with regional voice talent, capable of capturing the nuances of local accents and dialects from across the globe.
Which is why it also makes sense for software developers to do the same. Providing some personality to AI projects so that the software can better identify a dialect during its communications with users. This personality gives a higher potential for success and increased customer uptake.
Professional voice talent with a local touch can make all the difference to your software. Voice Talent Online has access to over 1,500 voice actors fluent in more than 75 languages. Contact us today to find out more.