Over the past few years, the internet has been bustling with revelations of everyday sexism that had previously been normalised or overlooked. As part of this, a debate arose as to why the vast majority of subservient AI (artificial intelligence) assistants are voiced by women. Could this stem from the deep-routed misogyny present in our society?
It was only in March 2021 that Apple decided to remove the original female voice of Siri as the default setting, allowing the user to choose their preference when initially setting up the phone. At the time, Apple was facing backlash about Siri’s default setting being this original female voice. Organisations such as UNESCO testified that this would prove to be detrimental to society in the near future. UNESCO stated that “voice assistants with female voices are perpetuating harmful gender biases.” They stated that engineers, who are mainly male, in these companies have created feminised AI that “greets verbal abuse with catch-me-if-you-can flirtation”.
As late as 2017, Siri would still respond to sexist or abusive language in a passive or humorous way, often playing into the stereotype that women should just accept verbal abuse, or even take it as a compliment. In recent updates, however, Siri has been programmed to react to rude language and sexist comments with stern answers such as “No”, or “I will not answer that”. This is a huge progression from the previous docile responses of “I’d blush if I could.”
In addition, the majority of the other ‘classic’ AI voices are female, such as Alexa, Cortana, and the default voice for Google Assistant.
These aspects seem to point towards the technology industry in being a deliberately sexist house of horror, run by misogynistic men whose only aim in life is to demeanour women. However, this conclusion may be premature. It is important, therefore, to consider why the female voice is the most used for these AI voice assistants, and how this might point to broader societal issues.
How do we respond to different voices?
For as long as public advertisement has been around, we have been able to see the divide between male and female voice preferences. Masculine voices are mainly used for serious information and are often seen as more authoritative or trustworthy. They are often used to sell cars or D.I.Y. equipment, cover sports games, and make government announcements.
However, according to a study done by The Voice Realm, 66% of internet video viewers prefer a female voice for a voiceover, finding it “friendlier”. When selling a product centred around hospitality, health, or education, female voices are often used to induce warmth and are seen as more persuasive than forceful.
Alexander Graham Bell hired a woman, Emma Nutt, as the first telephone operator in 1878, asserting that women were more polite, patient, and soothing, with her voice eventually becoming the archetype to which telephone operators were expected to conform.
Many studies prove that humans prefer the sound of the female voice. Suggesting that we adapted this from the womb, where the lull of our mother’s voice would set us at ease. Others state that women can articulate vowels more effectively than their male counterparts, making them easier to understand by a wider variety of people, and thus more pleasant to listen to.
Hope for the future?
Even though women only make up 22% of AI researchers, this number is increasing, suggesting that the industry is becoming more inclusive. This is already leading to changes in AI voicing, with Copenhagen Pride and a few other co-founders creating the first genderless AI, Q. The founders state that they are trying to “break down the gender binary” and to “break the mindset where female voices are preferred for assistive tasks and male voices for commanding tasks”.
Apple released the new iPhone 14 last month, with a revenue of $25.9 billion in September alone. These were all fitted with the new and improved IOS16 software update, which boasts 15 different voice options for Siri; 7 male and 8 female.
Amazon has also recently released their new voice “Ziggy” (inspired by its “A-to-Z” slogan), but only in the United States. I’m sure that die-hard David Bowie fans are desperately waiting for this feature to be introduced in the UK too so they can say “Hey Ziggy, play Rock ‘n’ Roll Suicide!”
These new releases, among others, are big steps in the right direction, and I predict advancements like this becoming more commonplace in the near future.
For now, we have highly intelligent AI voice assistants at our fingertips, something we cannot take for granted, regardless of their voices. Hey Siri, thanks for your help.