– OpenAI pauses use of Sky voice in ChatGPT over Scarlett Johansson likeness
– ChatGPT uses voices from paid voice actors, selected from 400 submissions
– GPT-4o model in ChatGPT will introduce multimodality and real-time conversation features
OpenAI has paused the use of the Sky voice in ChatGPT due to concerns that it sounds too much like Scarlett Johansson, the actress from the movie “Her.” The voices in ChatGPT were from paid voice actors, and the selection process for the voices was purely coincidental. OpenAI is rolling out a new GPT-4o model into ChatGPT, which will feature a new conversational interface with natural-sounding and emotion-mimicking AI.
The comparison to Scarlett Johansson’s character in “Her” became more apparent when OpenAI CEO Sam Altman and others noted the similarity between the new AI model and the movie. GPT-4o is different from previous models in that it is multimodal, trained on and capable of understanding images, text, video, and speech. This allows for real-time conversations and an emotional voice.
The five voices available for ChatGPT Voice were selected from over 400 submissions and licensed from voice actors who were compensated at top-of-market rates. OpenAI emphasized that the voices were not intentionally mimicking celebrities, including Scarlett Johansson, and each voice belongs to a different professional actress. There are concerns that the emotional and adaptable voice in ChatGPT Voice could be misused or used for deepfakes due to its realistic nature.
OpenAI has partnered with award-winning casting directors and voice actors, flying them to San Francisco for recording sessions to train the new AI voice models. Despite the similarities, the Sky voice is not an imitation of Scarlett Johansson, and OpenAI has not disclosed the name of the actress behind the voice. The movie “Her” is available to stream on Max.