OpenAI has announced the retirement of a ChatGPT voice called Sky due to its apparent similarity to Scarlett Johansson's voice. Despite the company's claims that the voice was not intentionally modeled after the famous actress, the decision was made to temporarily withdraw it after numerous media reports and jokes about the resemblance, including mentions on Saturday Night Live. OpenAI's CTO, Mira Murati, assured The Verge in an interview that the resemblance was not intentional.
An interesting controversy has arisen in the world of artificial intelligence - a ChatGPT voice called Sky has caused a stir due to its striking resemblance to the voice of Scarlett Johansson. Although OpenAI claimed the voice was the work of another professional actress, a number of complaints surfaced, forcing the company to withdraw the model.
OpenAI recently ran into unexpected trouble when their Sky voice model designed for ChatGPT came under fire for its similarity to Scarlett Johansson's voice. The model, which was introduced last year, gained a lot of attention and also criticism because it was said to resemble the star of the movie "Her" too much. Following the introduction of the new GPT-4o model, which includes more advanced voice features, the comparisons with Johansson have become even more pronounced.
Spike Jonze's 2013 film Her, in which Scarlett Johansson voices virtual assistant Samantha, inspired many discussions about how voice can make a human connection. In the film, Johansson showed how voice interpretation can express emotions and thoughts in a way that words alone cannot.
Despite OpenAI's assurances that the voice was not intentionally modeled after the famous actress, media outlets and even satirical shows like Saturday Night Live mocked the resemblance, leading to the company's decision to temporarily pull the voice. OpenAI also added that new voices are expected to appear in the future to better suit the diverse interests and preferences of users.
While OpenAI is not disclosing the names of the players who contributed their voices, they have confirmed that five voice profiles have been selected from over 400 applicants. The innovations brought by the GPT-4o model include more expressive voice interpretation, the ability to recognize facial expressions with the phone's camera and real-time translation of spoken language.
Conclusion: The controversy surrounding Sky's voice reveals the complexity of developing artificial intelligence that can mimic human voices and emotions. While OpenAI faces the challenges of ensuring authenticity and respecting privacy, it remains clear that the future of artificial intelligence will require even more sophisticated approaches and ethical considerations.
