ChatGPT creator OpenAI stopped using one of its AI-generated voices after some users said it had a flirtatious tone that sounded too much like Hollywood star Scarlett Johansson.
In a post on X, OpenAI said it was “working on pausing” Sky, the name of one of the five voices ChatGPT users can choose to speak with.
The company said it had “heard questions” about how it selects the realistic audio options available for its flagship AI chatbot, particularly Sky, and wanted to address them.
OpenAI was also quick to debunk internet theories about Johansson in an accompanying blog post detailing how the ChatGPT voices were chosen.
“We believe that AI voices should not deliberately imitate a celebrity’s distinctive voice: Sky’s voice is not an imitation of Scarlett Johansson, but rather belongs to a different professional actress using her own natural voice,” the company wrote.
Some users said she had a flirty tone that sounded too similar to that of Hollywood star Scarlett Johansson (pictured).
OpenAI said it was “working on pausing” Sky, the name of one of the five voices ChatGPT users can choose to speak with (File image)
Johansson starred in Spike Jonze’s 2014 film ‘Her’, in which she voiced an AI that Joaquin Phoenix’s character falls in love with.
ChatGPT users found some interactions with Sky to be strangely flirtatious. In a video posted by OpenAI, a female-voiced ChatGPT congratulates a company employee for “sporting an OpenAI hoodie.”
In another, the chatbot says “oh stop it, you’re making me blush” after being told it’s amazing.
“This is clearly programmed to feed guys’ egos,” Desi Lydic, the Daily Show’s chief correspondent, said in a segment last week. “You can really tell that a man built this technology.”
OpenAI first implemented voice capabilities for ChatGPT, which included five different voices, in September, allowing users to engage in conversations with the AI assistant.
‘Voice Mode’ was originally only available to paid subscribers, but in November, OpenAI announced that the feature would be free to all users with the mobile app.
Johansson starred in Spike Jonze’s 2014 film ‘Her’, in which she voiced an AI that Joaquin Phoenix’s character falls in love with (pictured)
OpenAI CEO Sam Altman seemed to take advantage of the Ella comparison: he simply posted the word “her” on social media platform
And ChatGPT interactions are becoming more sophisticated. Last week, OpenAI said the latest update to its generative AI model can mimic human cadences in its verbal responses and can even attempt to detect people’s moods.
OpenAI says the newest model, called GPT-4o, runs faster than previous versions and can reason through text, audio and video in real time.
In a demonstration during OpenAI’s announcement on May 13, the AI robot conversed in real time, adding emotion — specifically “more drama” — to its voice as requested.
It also attempted to extrapolate a person’s emotional state by watching a selfie video of their face, with the help of language translations, step-by-step math problems, and more.
GPT-4o, short for “omni,” is not yet widely available. It will progressively reach selected users in the coming weeks and months. The model’s text and image capabilities have already started rolling out and are set to reach even some of those using ChatGPT’s free tier, but the new voice mode will only be available to paid ChatGPT Plus subscribers.
While most are yet to get their hands on these newly announced features, the capabilities have evoked even more comparisons to Spike Jonze’s dystopian romance ‘Her’.
OpenAI CEO Sam Altman seemed to take advantage of this as well: he simply posted the word “she” on social media platform X on the day of the GPT-4o presentation.