OpenAI has announced its decision to withdraw one of the voices of the AI chatbot ChatGPT named Sky, after causing a stir due to its resemblance to the voice of actress Scarlett Johansson in the movie “Her,” which explores the topic of artificial intelligence.
The AI company, backed by Microsoft, made a statement on X platform: “We have heard concerns about how voices are chosen in ChatGPT, especially the voice named Sky. We have decided to temporarily suspend the use of Sky while we address these issues.”
The science fiction film “Her,” released in 2013, tells the story of a man who falls in love with an artificial intelligence system named Samantha, which used the voice of actress Scarlett Johansson.
These developments come just one week after OpenAI launched a set of voices for the ChatGPT chatbot, a new artificial intelligence model named GPT-4o, and the computer versions of ChatGPT.
Viewers who watched the live stream of ChatGPT’s voice capabilities immediately shared videos on social media claiming that Sky’s voice resembles Johansson’s.
On the day President Joe Biden said that artificial intelligence could pose a national security threat, a post on X platform said.
In a blog post, OpenAI announced that they selected five voices for the chatbot through a selection and recording process that lasted five months.
The company contracted experts to choose voice actors by narrowing down participants from 400 to 14, and then the internal team selected the final five.
OpenAI has stated that it does not believe it is necessary for artificial intelligence voices to intentionally mimic a celebrity’s voice, and Sky’s voice is not an imitation of Scarlett Johansson’s voice even though a different professional actress’ voice was used. The names of their voice talents have been withheld to protect their privacy.
OpenAI does not currently clarify the reason for removing Sky’s voice or the steps to be taken before reintroducing it in ChatGPT.
The company plans to conduct tests for the voice mode feature in the coming weeks, where ChatGPT Plus subscribers will have early access to this feature, in addition to the company’s plans to add new voices as well.
OpenAI also confirmed that the new model can respond to user requests in less than 232 milliseconds, with an average response time of 320 milliseconds, similar to human response time in conversations.
Founded in 2015, the company is valued at over $80 billion. OpenAI faces pressure to lead in the generative AI market by finding ways to generate profits, as it spends large sums on processing and infrastructure to build and train its models.
OpenAI, Microsoft, and Google lead the artificial intelligence field, with companies competing in all industries to add chatbots and agents that use AI to avoid falling behind competitors in the field.
