The Evolving Voices of Artificial Intelligence

The Evolving Voices of Artificial Intelligence
The Evolving Voices of Artificial Intelligence

What does AI sound like? Hollywood has been thinking about this for a long time, and now AI developers are following suit, creating voices for real machines inspired by movie adaptations.

OpenAI recently introduced updates to its AI chatbot, ChatGPT, which now boasts the ability to hear, see, and speak in a natural voice. This voice, called Sky, bears a striking resemblance to the operating system voiced by Scarlett Johansson in the 2013 film “Her.” With a husky, soothing tone, Sky’s voice was designed to be pleasant and versatile. However, Johansson has expressed concerns about the similarities, having previously declined to lend her voice to the project. OpenAI claimed that Sky was voiced by another actress, but agreed to discontinue its use out of respect for Johansson. Users have since called for Sky’s return.

AI developers often emphasize the naturalistic qualities of their creations, but these voices are steeped in layers of artifice and projection. Sky exemplifies OpenAI’s cutting-edge ambitions, but she also reflects an outdated notion of AI as an empathetic and compliant female presence. Such stereotypes, present in characters like Samantha from “Her,” continue to be perpetuated in AI technology.

Historically, female voices have been integral to the representation of fictional technologies before their implementation in the real world. In the original «Star Trek» series, the computer’s voice was provided by Majel Barrett, while in the 1979 film «Alien,» the ship’s computer was called «Mother.» When tech companies began releasing virtual assistants like Apple’s Siri and Amazon’s Alexa, these voices were predominantly female, reinforcing certain gender expectations.

These early voice assistants, characterized by their mechanical, monotonous tones, evoke a sense of nostalgia rather than realism. Despite technological advances, artificial sound persists, in part because of its programmable and subordinate connotations, which make users feel in control.

Text-to-speech software, originally designed for accessibility, has found creative use on platforms like TikTok, where various simulated voices are available. However, the most popular voice, “Jessie,” embodies a lively, robotic tone, often used to add a touch of humor to videos.

Male AI voices also have their place in cinematic history, with HAL 9000 from «2001: A Space Odyssey» being a prime example. HAL’s calm and loyal demeanor becomes chilling when he challenges the protagonist, demonstrating how male voices can be portrayed as both allies and adversaries.

In «Her,» Samantha goes from a simple assistant to a living being, a narrative that complicates the typical portrayal of AI voices. Johansson’s performance as Samantha feels genuinely human, not just in her voice, but in the character’s development throughout the film.

OpenAI’s Sky, inspired by Johansson’s Samantha, aimed to project an advanced, sentient quality. However, revisiting “Her,” it’s clear that such AI voices are simulations built on layers of previous representations, devoid of true human qualities.

Tech companies market their virtual assistants based on their functionality and their ability to make users feel comfortable with AI. Johansson’s voice, seen as a comforting bridge between technology and users, highlights this trend. OpenAI strives for “accessible” and “trustworthy” voices, aiming to ease the discomfort of interactions with AI.

Last September, OpenAI introduced Sky’s voice alongside other options like Juniper, Ember, Cove, and Breeze. When interacting with these voices, users often find themselves perceiving them as more human than they actually are. Despite the naturalistic touches, these voices remain fundamentally artificial, designed to perform specific roles rather than displaying true intelligence.

Sky’s temporary removal has led to user dissatisfaction, with many expressing a desire for a new female voice with a different personality. This ongoing evolution of AI voices continues to reflect and shape our interactions with technology, balancing familiarity with unease.

Produced by Safie’s Tale

Audio via Warner Bros. (Samantha, HAL 9000); OpenAI (Sky); Paramount Pictures (Enterprise Computer); Apple (Siri); TikTok (Jessie)