Audio by Jarvis H. using WellSaid Labs
The choices we make in pursuit of new ways of thinking and doing often prompt us to re-evaluate what we already thought we knew.
One way we’ve seen this play out at WellSaid is in the way we think about the relationship between the visual and the audio. Imagine for a moment that you had to select a stock image to represent what your voice sounds like. What would my voice look like if not like, well, me?
Recently, we revisited a choice we made as an early company that faced this very practical challenge. Here is a story of one of the ways WellSaid has stayed rooted in our values as we grow.
How We Created Our Original Avatars
In our recent blog post, we described the steps we take to protect the identities of WellSaid voice actors. This responsibility has been a priority for us since our early days as a company.
In the early days, we wrestled with the question of how to present our synthetic voices to our customers – do they look like people? Do they have first and last names? Are they simply play icons? We wanted the voices to feel and sound as approachable and natural as possible while also preserving the anonymity of the voice actors who provided the data for the voices.
To address this challenge, we asked our team members who knew nothing about the identities of the voice actors to listen to original recordings for our synthetic voices. As they listened, they were asked, what personality do they hear in this recording? How would you describe the person you are hearing?
Once the team members had a sense of a personality for that synthetic voice, they selected a photo from a pool of stock images to represent them visually. This is the process we followed to create our 14 original voice avatars.
Our Commitment
WellSaid’s foundational values include building with intent and holding ourselves accountable for what our technology enables. As our company matured, so did our understanding of how we can put our values to work in our partnership with voice actors.
While our synthetic voices are not living beings, and do not have a race or a gender, these voices are built using data from human voice actors who do have racial and gender identities. Voice talent of color have experienced and continue to experience underrepresentation and misrepresentation in the voiceover industry in the US and beyond.
We are committed to avoiding these harmful practices in our work with voice actors.
Putting Our Values to Work
In 2021, we began to ask all new actors who recorded with us to choose a stock image and name for the voice we built with their recordings. We continue this process with any new voice talent to this day.
To align our practices, we recently asked our 14 original voice talent to also select avatar images that more closely align to their identities. We believe this approach better reflects our commitment to holding ourselves accountable for the biases in our processes, and to actively minimize potential harm.
What this Change Means For Our Customers
The next time you log into Studio, you might see new images for some of our avatars. Rest assured, their names haven’t changed, and you’ll still be able to find and use their amazing voices!
If you have any questions, please reach out to support@wellsaidlabs.com.
