OpenAI says it plans to stop using certainly one of its ChatGPT voices after some users said she appeared like Scarlett Johansson, who famously voiced a fictional and then-futuristic AI assistant within the 2013 film “Her.”
In one post On the social media platform the corporate said it had “heard questions” about the way it selects the lifelike audio options available for its flagship artificial intelligence chatbot, particularly Sky, and desired to answer them.
OpenAI also quickly debunked the Internet’s theories about Johansson in an accompanying blog post that detailed how ChatGPT’s votes were chosen.
“We believe that AI voices should not intentionally mimic a celebrity’s distinctive voice – Sky’s voice is not an imitation of Scarlett Johansson, but belongs to another professional actress using her own natural speaking voice,” the corporate said wrote. For data protection reasons, the corporate shouldn’t be allowed to pass on the names of its voice actors.
San Francisco-based OpenAI didn’t comment further on why it decided to pause its use of Sky.
OpenAI first introduced voice capabilities for ChatGPT in September, which included five different voices and allowed users to converse with one another using the AI assistant. Voice Mode was originally only available to paid subscribers, but OpenAI launched in November announced that the feature will change into free for all mobile app users.
And ChatGPT’s interactions have gotten increasingly sophisticated. Last week OpenAI said this latest update to its generative AI model can mimic human rhythms in his verbal reactions and even try and recognize people’s moods.
According to OpenAI, the most recent model, called GPT-4o, works faster than previous versions and might process text, audio and video in real time. In an indication through the OpenAI announcement on May 13, the AI bot chatted in real time and added emotions – specifically “more drama” – to its voice as requested. It was also an try and extrapolate an individual’s emotional state by watching a selfie video of their face, which helped with language translations, step-by-step math problems and more.
GPT-4o, short for “Omni,” shouldn’t be yet generally available. It will progressively make its technique to select users over the approaching weeks and months. Rollout of the model’s text and image features has already begun and is predicted to achieve even a few of those on ChatGPT’s free tier – nonetheless, the brand new voice mode will only be available to paid ChatGPT Plus subscribers.
While most have not gotten their hands on these newly announced features yet, the features have drawn much more comparisons to Spike Jonze’s dystopian romance “Her,” about an introverted man (Joaquin Phoenix) who falls for an AI Operator falls in love with System (Johansson), which results in many complications.
OpenAI CEO Sam Altman also appeared to benefit from this – quite simply publication the word “they” on social media platform X on the day of GPT-4o’s unveiling.
Many who responded to the model’s demos last week also found that among the interactions took on an oddly flirtatious tone. In one Video For example, posted by OpenAI, a female-voiced ChatGPT praises an organization worker for “rocking an OpenAI hoodie,” and in one other chatbot says “Oh stop, you’re making me blush” after being told it was amazing.
This has sparked some discussion concerning the gendered ways in which critics say tech firms have long used to develop and interact with voice assistants, long before the recent wave of generative AI advanced the capabilities of AI chatbots. In 2019, the United Nations cultural and scientific organization was founded pointed to “firmly anchored submissiveness.” built into standard female-voiced assistants (from Apple’s Siri to Amazon’s Alexa), even when faced with sexist insults and harassment.
“This is clearly programmed to feed guys’ egos,” Daily Show chief correspondent Desi Lydic said in a segment about GPT-4o last week. “You can really tell that a man developed this technology.”
