AI brands continue to promise more uniform and human-like experiences for their products, and Microsoft is no different. The brand is working to make the features of its AI assistant more cohesive with the everyday world. This includes a more authentic Spanish language voice for its Copilot product. The company rolled out its Copilot Voice feature last October, which it will soon update with native Spanish speaking-voices to cater to the Latino community in the U.S. and globally.
Enter Elm and Alder, Copilot’s native Spanish voices. The AI can detect when Spanish is being spoken and respond seamlessly in fluent Spanish with a native accent. Similarly, it expresses a native accent while speaking English in the Spanish setting. I spoke with a Microsoft product expert to explore the process behind developing the new voices and the intent behind providing this unique language and cultural representation within Microsoft’s AI suite.
“In the U.S. market, you have all of these varied and diverse backgrounds and people that are your audience. In our research, we want to be representative with our communities,” says Microsoft AI Product Designer, Laura Parra Rangel. “We noticed that there was a bit of a gap for the Hispanic community in the U.S. There’s a huge Latinx community in the U.S. We have a huge community that can benefit from this.”
The impact of Spanish voices
Copilot Voice has already received some positive reviews in its English version, with PCMag’s Michael Muchmore describing an early experience with the AI voice feature as “almost like talking with a friend.” He noted that it will still respond even amid normal speech patterns, such as stumbling with speech or interrupting the AI. Copilot will continue to fulfill its query.
Such a feature would surely be beneficial when implemented into additional languages in a native fashion. Particularly, for a demographic that is collectively the second-largest ethnic group in the U.S. According to Census.gov, over 65 million people identified their ethnicity as Hispanic or Latino in 2023.
“For me, the voice experiences are amazing. Now that we have Spanish options, it’s just a game changer for me– an actual decent Spanish option to have these experiences. Something that can change and switch between English and Spanish, just like I do in my regular life,” Parra Rangel says.
“I’m talking about something in English and then something came to my mind in Spanish, and I’m constantly switching between one and the other, and I’m sure this is true for many, many people across the world,” she added.
In conjunction with the coming Copilot native Spanish language rollout, Microsoft released a report surveying 504 U.S. Latina moms ages 25 to 40 on how they use AI in their everyday lives. The campaign, called “Abuela Approved AI” found that 56% of Latina moms consider AI as a part of their close family support team; 60% of moms include AI in their daily tasks, including entertainment, personal growth, work or business, and language translation; 70% of moms note that they use AI for 24/7 access to support over advice from others, while 51% indicate they consider AI to be a source they can access for non-judgmental support. Finally, 87% of Latina moms referenced AI for cultural enrichment and preservation, including hobbies, family recipes, and native language, with AI being a bridge for learning old concepts and new ones.
Some practical examples Latina mothers gave for how Copilot Voice has benefitted them in their daily lives include using the mobile app to scan available ingredients in their homes to create a recipe that aligns with their culture. They also detailed using prompts in Spanish to create unique, first-person, native language, bedtime stories to promote bilingualism and multiculturalism in their children.
“I speak Spanish fluently, it’s my first language. English is my second language, but I totally see people being first generation in the U.S. being like– I want to speak Spanish. I want to preserve my culture. And having Copilot helps them preserve their culture,” she says.
“It’s great to have a true Hispanic voice in the market that I can speak to and that sounds [familiar]. With an Argentinian accent or with a Dominican Republic accent, that’s amazing,” she added.
Voice on all platforms
Copilot Voice first launched using OpenAI’s GPT-4o AI model. Other services such as Microsoft’s Think Deeper also include Copilot Voice and use more powerful models, such as GPT o1. Notably, a feature like Think Deeper is primarily web-based; however, users can access other Copilot Voice setups via Microsoft Windows and macOS desktops, and iOS and Android mobile interfaces.
Parra Rangel noted that working with the Microsoft AI team changed her perspective on how voice features could be presented. Having previously associated voice strictly with mobile, observing the development of Copilot Voice allowed her to see how a voice feature can also be compatible with a desktop experience.
“Of course, you always have the option to use voice in desktop. As a product designer, I used to see that as only mobile, but now with this, it’s a game-changer. How we experiment with voice on all platforms, whether it’s computer or mobile. This is just the tip of the iceberg,” she says.
I asked Microsoft whether there would be any model updates for the Copilot native Spanish feature and await an answer. I also asked the company whether it has plans to release native versions of other languages to Copilot Voice, and it said it will follow up with an answer.
Copilot’s wider appeal
While tech brands are promoting the highlights of new features, the product manager acknowledged users’ potential concerns about trusting AI, including privacy, security, and other notable errors or glitches. She noted that some people might resist using Copilot due to the stigma of using AI decreasing intelligence, or feeling like AI might take away their ability to think or create originally. Parra Rangel said users can think of Copilot as someone to bounce ideas with or to ask a quick question, not to generate content fully.
“Something that Copilot has is that it is aimed for all age groups. It’s unlike a more niche social media that’s aimed for maybe Gen Z, or only Millennials. I feel that Copilot is super inclusive in that sense. No matter your age, you can always benefit from Copilot,” she says.
Giving an example of a personal way she uses Copilot, Parra Rangel noted she takes advantage of the seamless language settings to create an in-app language learning tool that she can use on the go. There is a speech rate setting that allows users to adjust how fast or slow the voice speaks. She said she will slow the voice for better understanding and use Copilot Voice as her personal coach in language training. Currently, Copilot supports 40 languages, and 14 are considered very well supported.
Parra Rangel considers that the introduction of native Spanish voices has the potential to promote language learning for Copilot users regardless of whether they are Hispanic.
“You can still benefit from having these truly, Spanish-speaking voices for learning new things, a new language, or if you want to improve your Spanish, or if you live in a diverse area where people speak Spanish or other languages. That’s a very nice way to get to learn your neighbor’s culture and language,” she says.
She added that the only way to dispel any hearsay about Copilot and its features is to give it a try.
