Skip to main content

From Sirens to Siri

Francesca Lawson explains the historical biases behind female singing.


“Hey Siri, can you sing for me?”

In response to this question, Siri provided the lyrics of “If I Only Had a Brain” via text, but Siri didn’t say anything, let alone sing. iPhone-owners have been trying to get Siri to sing since her invention. But year after year, while Apple rolls out update after update, one thing remains the same: Siri cannot sing.

Addressing Siri’s lack of musical ability, Francesca Lawson (Ethnomusicology) gave a presentation on March 2, 2023, for the Humanities Center Colloquium called “Why Can’t Siri Sing? Cultural Narratives That Constrain Female Singing Voices in AI.” Lawson’s presentation addressed several points, including how our Western culture is infused with both fear of and fascination with the female singing voice in history and in modern AI.

Historical Cultural Narratives

Lawson set the stage for her analysis of Siri by providing several thousand years of contextual history of how women’s voices have been traditionally viewed in Western culture. The fear that the feminine voice (especially the singing voice) could be the undoing of men has permeated the Western experience, starting with sirens in ancient mythology. During the medieval era and the Renaissance, Eve was seen as a femme fatale figure; as a result, the female voice was not permitted in churches or on stage. And yet, there was a desire for clear soprano voices, leading men to prefer mutilating boys so that their voices would sound like female voices (castrati) rather than allowing a woman on stage. When laws changed and women began to permeate opera, fear and fascination of “the diva” (prima donna) reigned. This caused the female voice to be even more sexualized because it was allowed openly. This sexualization has continued all the way to pop stars today.

Lawson gave two aspects of interactions with the female voice that, by her research, demonstrate long-standing misogyny. She said that even in the case of AI, the female voice is faced with ridicule and sexualization. “Siri’s nature is that of a servant that has no choice but to be helpful,” Lawson said. Siri serves as both domestic servant and personal assistant. While there are other voice options for Siri, the dominant voice is feminized and programmed for servitude, programmed to be incapable of responding to ridicule—reinforcing female acceptance of ridicule and mockery on the public stage.

Siri is often met with such treatment. Lawson gave the example of the short time the “Bohemian Rhapsody” trick was part of Siri’s programming. Videos on YouTube mock Siri’s flat, robotic recitation of the slightly modified lyrics by rock band Queen. Rather than improve Siri’s ability to sing, Apple programmers removed the feature altogether. In another example, a YouTuber programmed a rap battle between Siri and Alexa as part of the “cat fight” trope, which Lawson said, “ridicules and disparages women” since Siri and Alexa roasted each other with songless rhymes and a background beat. “The effect of ridicule is to nullify women aspiring to power.”

AI That Can Sing

While Siri has never been given the gift of song, there is plenty of AI technology that can, in fact, sing. Lawson introduced her audience to one of them. Sophia the Robot is a robot designed to look humanoid and actually is programmed with the ability to sing. And, to Lawson’s point, the reality of it is. . . chilling.

Lawson explained that chills typically accompany a “peak musical listening experience” and also a biological survival reaction. Chills accompany the experiences of both fascination and fear. In the case of listening to “robotic sirens,” we perceive the voices as both desirable and threatening.

After showing a video of a duet between Sophia and Jimmy Fallon, Lawson shared the extreme user reactions posted in the video comments section. Listeners wrote about their anxiety and fear, how creepy they thought her singing was. Many wrote how the duet gave them the “chills” Lawson discussed.

Lawson concluded, “The problem of the female singing voice has been with us for millennia. If it isn’t intentional, it is so much a part of our thinking that it comes out in these ways.” Robots and AI are becoming a larger part of society’s interactions with the world. According to Lawson, “examining the underlying historical blueprints” to how we respond to things will help us improve connectivity and the battle between technological expectation and emotional reaction to the female voice.