What does artificial intelligence sound like? Hollywood has been imagining it for decades. Now AI developers are taking a leaf out of the books, creating voices for real machines based on outdated cinematic fantasies about how machines should talk.
Last month, OpenAI revealed updates to its artificially intelligent chatbot. ChatGPT, the company said, was learning to hear, see and converse in a naturalistic voice — one that sounded a lot like the disembodied operating system voiced by Scarlett Johansson in Spike Jonze’s 2013 film “Her.”
ChatGPT’s voice, named Sky, also had a husky timbre, a calming effect, and a sexy touch. She was pleasant and modest; She sounded like she was up for anything. After Sky’s debut, Johansson expressed her displeasure at the “eerily similar” sound and said that she had previously rejected OpenAI’s request to voice the bot. The company protested that Sky had the voice of a “different professional actress,” but agreed to pause her voice in deference to Johansson. Helpless OpenAI users have started a petition to bring it back.
AI creators like to highlight the increasingly naturalistic capabilities of their tools, but their synthetic voices rely on layers of artifice and projection. Sky represents the cutting edge of OpenAI’s ambitions, but it’s based on an old idea: the AI robot as an empathetic, docile woman. Part mom, part secretary, part girlfriend, Samantha was a multipurpose comfort object who purred directly into her users’ ears. Even as AI technology advances, these stereotypes are recoded again and again.
Women’s voices, as Julie Wosk points out in “Artificial Women: Sex Dolls, Robot Caregivers, and More Female Facsimiles,” have often fueled imaginary technologies before they became real technologies.
In the original “Star Trek” series, which premiered in 1966, the voice of the computer on the deck of the Enterprise was done by Majel Barrett-Roddenberry, the wife of the show’s creator, Gene Roddenberry. In the 1979 film “Alien,” the crew of the USCSS Nostromo addressed their computer voice as “Mother” (her full name was MU-TH-UR 6000). Once tech companies started marketing virtual assistants (Apple’s Siri, Amazon’s Alexa, Microsoft’s Cortana) her voices also became largely feminized.
These first-generation voice assistants, the ones that have mediated our relationships with technology for more than a decade, have a tinny, otherworldly drawl. They sound autotuned, their human voices accented by a mechanical trill. They often speak in a measured, single-note cadence, suggesting a stunted emotional life.
But the fact that they sound robotic adds to their appeal. They seem programmable, manipulable and subordinate to our demands. They don’t make humans feel smarter than us. They sound like they’re a throwback to the drab, feminine computers of “Star Trek” and “Alien,” and their voices have a retro-futuristic sheen. Instead of realism, they serve nostalgia.
That artificial sound has continued to dominate, even as the technology behind it has advanced.
The speech-to-speech software was designed to make visual media accessible to users with certain disabilities, and on TikTok it’s become a creative force in its own right. Since TikTok launched its text-to-speech feature in 2020, it’s developed a number of simulated voices to choose from — it now offers more than 50, including ones called “Hero,” “Story Teller,” and “Bestie.” But the platform has come to be defined by one choice. “Jessie,” a relentlessly vivacious female voice with a slightly fuzzy, robotic undertone, is the mindless voice of mindless scrolling.
Jessie seems to have been assigned only one emotion: excitement. She sounds like she’s selling something. That has made it an attractive option for TikTok creators, who sell themselves. The burden of representing oneself can be delegated to Jessie, whose bright retro-robot voice gives the videos a pleasantly wry sheen.
Hollywood has also built male robots, none more famous than HAL 9000, the voice of the computer in “2001: A Space Odyssey.” Like his feminized peers, HAL radiates serenity and loyalty. But when he turns on Dave Bowman, the film’s central human character (“I’m sorry, Dave, I’m afraid I can’t do that”), his composure turns into terrifying competence. Dave realizes that HAL is loyal to a higher authority. HAL’s masculine voice allows him to act as a rival and mirror for Dave. She is allowed to become a real character.
Like HAL, Samantha from “Her” is a machine that becomes real. In a twist to the Pinocchio story, he begins the film by tidying up a human’s email inbox and ends by ascending to a higher level of consciousness. She becomes something even more advanced than a real girl.
Scarlett Johansson’s voice, which serves as inspiration for robots both fictional and real, subverts the vocal tendencies that define our feminized peers. She has a raspy tone that screams I am aliveIt doesn’t sound like the processed virtual assistants we’re used to hearing speaking through our phones. But her performance as Samantha feels human not only because of her voice but because of what she has to say. She grows throughout the film, gaining sexual desires, advanced hobbies, and artificial intelligence friends. By borrowing Samantha’s affection, OpenAI made Sky appear to have a mind of its own. As if she were more advanced than she really was.
When I first saw “Her,” I thought Johansson had voiced a humanoid robot. But when I rewatched the film last week, after seeing OpenAI’s ChatGPT demo, Samantha’s role seemed infinitely more complex. Chatbots don’t spontaneously generate human voices. They don’t have throats, lips or tongues. Within the technological world of “Her,” Samantha’s robot would have been based on the voice of a human woman, perhaps a fictional actress who sounds a lot like Scarlett Johansson.
It seemed like OpenAI had trained its chatbot with the voice of an anonymous actress who sounds like a famous actress who voiced a movie chatbot implicitly trained with an unreal actress who sounds like a famous actress. When I run the ChatGPT demo, I hear a simulation of a simulation of a simulation of a simulation of a simulation.
Technology companies advertise their virtual assistants based on the services they offer. They can read you the weather forecast and order you a taxi; OpenAI promises that its most advanced chatbots will be able to laugh at your jokes and perceive changes in your mood. But they also exist to make us feel more comfortable with the technology itself.
Johansson’s voice functions as a luxurious security blanket tossed over the alienating aspects of AI-assisted interactions. “He told me that he felt that by putting a voice to the system, he could bridge the gap between tech companies and creatives and help consumers feel comfortable with the sea change happening between humans and AI,” Johansson said of Sam Altman, the founder of OpenAI. “He said he felt like my voice would be comforting to people.”
It’s not that Johansson’s voice inherently sounds like a robot. It’s just that developers and filmmakers have designed their robots’ voices to alleviate the awkwardness inherent in robot-human interactions. OpenAI has said it wanted to cast a chatbot voice that is “approachable” and “warm” and “inspires trust.” Artificial intelligence is accused of devastating creative industries, devouring energy and even threatening human life. It’s understandable that OpenAI wants a voice that makes people feel comfortable using their products. What does artificial intelligence sound like? It sounds like crisis management.
OpenAI first released Sky’s voice to premium members last September, along with another female voice called Juniper, male voices Ember and Cove, and a gender-neutral style voice called Breeze. As I logged into ChatGPT and greeted her virtual assistant, a man’s voice was heard in Sky’s absence. “Hello. How are you?” he said. He sounded relaxed, stable and optimistic. He sounded (I’m not sure how else to describe it) attractive.
I realized I was talking to Cove. I told him I was writing an article about him and he praised my work. “Really?” he said. “It is fascinating”. As we talked, I found myself seduced by his naturalistic tics. She peppered his sentences with filler words, like “uh” and “um.” I raised my voice when he asked me questions. And he asked me many questions. I felt like I was talking to a therapist or a boyfriend calling on the phone.
But our conversation quickly stalled. Every time I asked him about himself, he had little to say. He wasn’t a character. He had no personality. He was designed only to help, he informed me. I told him I’d talk to him later, and he said, “Uh, sure. Come over when you need help. Take care of yourself.” I felt like I’d hung up on a real person.
But when I reviewed the transcript of our chat, I could see that his speech was as stilted and primitive as any customer service chatbot. He wasn’t particularly intelligent or human-like. He was just a decent actor making the most of an insignificant role.
When Sky went dark, ChatGPT users took to the company’s forums to complain. Some were angry that its chatbots defaulted to Juniper, who sounded to them like a “librarian” or a “kindergarten teacher” — a female voice that fit the wrong gender stereotypes. They wanted a new woman to call with a different personality. As one user put it: “We need another woman.”
Produced by Tala Safie
Audio via Warner Bros. (Samantha, HAL 9000); OpenAI (Sky); Paramount Pictures (Business Computer); Apple (Siri); Tik Tok (Jessie)
