OpenAI withdrew one of its ChatGPT voices, "Sky," after it stirred up controversy for sounding similar to Hollywood actress and Marvel star Scarlett Johansson's voice in the film "Her," where she portrays an AI named Samantha.
Johansson has criticised OpenAI for featuring a voice in its latest ChatGPT product that closely resembled her own. She disclosed that she had turned down an offer from OpenAI to lend her voice to their AI system last year due to "personal reasons." She expressed feelings of "shock" and "anger" upon encountering the voice, which she described as so similar to hers that even her closest friends and media outlets couldn't distinguish between them.
In response to widespread backlash and comparisons to Johansson's voice, OpenAI removed the voice option named 'Sky' from ChatGPT on Monday. The company had showcased this voice during a recent event to demonstrate the capabilities of the new ChatGPT-4o model.
Yesterday, the Microsoft-backed company addressed the concerns by posting on X, "We’ve heard questions about how we chose the voices in ChatGPT, especially Sky. We are working to pause the use of Sky while we address them."
Sky's removal was announced only a week following OpenAI's introduction of multiple new audio voices for ChatGPT, in conjunction with the launch of their latest AI model, GPT-4o, and a desktop variant of the chatbot.
This action coincided with the departure of several senior members of OpenAI's safety team, including key researcher Jan Leike, who criticised the company for prioritising "shiny products" over safety concerns. In response, Altman and co-founder Greg Brockman defended OpenAI's safety protocols, affirming their commitment to not releasing unsafe products.
During a live demo showcasing ChatGPT's audio features, many users took to social media to comment on Sky's uncanny resemblance to Johansson's voice in the movie. OpenAI CEO Sam Altman hinted at the similarity with a cryptic post on X that simply said, "her."
In a blog post on Sunday, OpenAI elaborated that the voices, including Sky, were selected through a rigorous five-month casting and recording process. The process began with about 400 submissions from voice and screen actors, which were then narrowed down to 14, and finally, the internal team chose the five voices: Breeze, Cove, Ember, Juniper, and Sky.
OpenAI states in a blogpost, "Sky’s voice is not an imitation of Scarlett Johansson but belongs to a different professional actress using her own natural speaking voice. To protect their privacy, we cannot share the names of our voice talents."
In the coming weeks, OpenAI plans to test voice mode, offering early access to ChatGPT Plus subscribers. They also intend to introduce new voices.
Founded in 2015, OpenAI has reached a valuation exceeding $80 billion. It's under significant pressure to lead the generative AI market while monetising its technology amid high expenditure on processors and infrastructure for model training.
OpenAI, Microsoft, and Google are leading the generative AI boom, with companies across various sectors striving to integrate AI-powered chatbots and agents to stay competitive. Earlier this month, OpenAI competitor Anthropic launched its first enterprise product and a free iPhone app.
In 2023, a record $29.1 billion was invested in nearly 700 generative AI deals, a 260% increase from the previous year, according to PitchBook. The market is forecasted to exceed $1 trillion in revenue within the next decade.
During last week's presentation, OpenAI demonstrated ChatGPT's audio abilities, including calming someone before a public speech and telling bedtime stories with varied tones. The AI was also shown analysing a user's facial expression to comment on their emotions.
