Siri, Alexa, Cortana: When AI speaks, why always a woman's voice?

Recently it emerged that Amazon has a team of employees who listen to some of the things we say to devices powered by the Alexa digital assistant. — dpa

The big three are all female. Apple's AI is called Siri, a Scandinavian female name. Amazon's smart assistant isn't Alexander, it's called Alexa. Google's may not have a name, but it's voice certainly is feminine.

"Voice assistants are the maidservants of our time," remarks Holger Schulz, a professor of musicology and head of the Sound Studies Lab at the University of Copenhagen, in a recent segment on the culture programme of the German public radio station Deutschlandfunk. "The digital maidservants in today's world solidify and perpetuate gender hierarchies, unequal treatment and exploitation."

Siri, the voice assistant in Apple devices such as the iPhone, sounds distinctly female. But asked whether it's male or female, the voice answers: "Don't let my voice fool you: I don't have a gender."

Alexa, on the other hand, says "she" considers "herself" female. The Google Assistant, while having no gender-specific name, has a female voice too.

Apple won't say how it comes up with the names of its products. But it points out that, depending on the language, there are Siri voice options for gender and dialect: "Some languages might offer only a male or female voice, while others offer both."

An Amazon spokesman says Alexa was named after the ancient Egyptian Library of Alexandria, regarded for centuries as the world's main centre of scholarship. Regarding the voice, the spokesman notes: "Studies all over the world have shown that the female voice is perceived as more pleasant, friendly and relaxed (than the male)." 

As Microsoft puts it, Cortana is first and foremost a dialogue-oriented assistant that supports users with different needs. It's not specifically female, but more generally a "being."

According to Microsoft, its "personality" is adapted to national conditions, so the French Cortana is different than the Japanese one, for example.

Miriam Meckel, a professor of corporate communication at the University of St Gallen in Switzerland, can also point to a number of studies showing that female voices are widely perceived as more pleasant than male ones. This is why the default setting for all voice assistants is female, she says.

The problem, Meckel argues, is that these voice assistants perform a service that is then associated with women in the real world. "Since more and more children are also growing up interacting with Alexa and the others, this can have an influence on society's perception of gender roles," she says.

Verbal sexual harassment is even directed at the female voice assistants, who "get quite an earful indeed," Meckel remarks. The tech companies are consequently keen to make sure that the responses are apt and not unduly polite.

"If you make Siri an indecent proposal today, you'll hear: 'The answer is no,'" she says. This shows, she goes on, that "rules of conduct are also important in communication and interaction with software systems. After all, we don't only learn from people now, but from machines as well.

"Modern voice assistance systems use real voices from which sentences are synthesised, and female voices have traditionally been used in technology and pop culture. If you call the talking clock in Germany, for instance, you'll still hear a woman.

In the 2013 US movie Her, a man falls in love with Samantha, an artificially intelligent virtual assistant – voiced by Scarlett Johansson – in a computer operating system. 

Judith Meinschaefer, a professor of Gallo-Romance linguistics at the Free University of Berlin, explains the advantage of female voices from a scientific standpoint: "Women's voices are more easily understood than men's," she says, because their vocal cords vibrate faster, resulting in sounds with a higher resonant frequency.

Technically, however, there would be no problem in using male voices, Meinschaefer adds. The likely reason that female voices are the norm, she suggests, is because the computer systems are predominantly designed by men, whose worldview may see women as assistants.

Things could have turned out differently though. In a YouTube video of a US television show in 1992, then Apple CEO John Sculley and computer researcher Kai-Fu Lee are shown presenting a prototype voice recognition system able to talk back to the computer user.

Its voice was rather high-pitched, but the system's name was Casper. – dpa