A new flagship model from a tech business is described as “the future of interaction between ourselves and the machines.”
OpenAI declared on Monday that it was releasing changes that included a new desktop service and improvements to its voice assistant features, along with the launch of its new flagship AI model, named GPT-4o. A delighted audience in the OpenAI offices saw Chief Technology Officer Mira Murati take the stage and promote the new model as an advancement in artificial intelligence. The faster and more accurate GPT-4 AI model, which was previously only available to paying clients, will now be available to free users thanks to the new model. “We’re examining how humans and machines will interact in the future,” Murati stated. “We believe GPT-4o is significantly changing that paradigm.”
A few of the changes that OpenAI unveiled on Monday were enhanced quality and speed of ChatGPT’s foreign language skills, along with the option to contribute text documents, audio files, and photos for the model to evaluate. In order to make sure features were utilized securely, the business said that it would be releasing them gradually.
A live demonstration of the model’s enhanced voice skills was also presented during the event, featuring two OpenAI research leads conversing with an AI voice model. Researchers told the voice assistant to talk with a range of different emotional and vocal inflections, and it came up with a bedtime narrative about love and robotics. Another example showed the AI model a math equation via the phone’s camera feature, and ChatGPT’s speech mode guided them through the solution. During the demonstration, an AI model was asked by a researcher to interpret the subject’s facial expressions and determine their emotional state. He appeared to be “happy and cheerful with a big smile and maybe even a touch of excitement,” according to ChatGPT’s voice assistant.
“You seem to be in a great mood, whatever’s going on,” ChatGPT stated in a lively, feminine voice. “Want to share where those positive energy came from?”
Last week, there were rumors that OpenAI would be introducing a search tool to take on Google, but according to Reuters, the business has postponed revealing that initiative. CEO Sam Altman tweeted on Friday, saying, “We’ve been hard at work on some new stuff we think people will love!,” but he disputed that Monday’s announcement was a search engine in response to the reports. seems magical to me.
Most of the demonstration proceeded without a hitch, with the exception of a few minor bugs or unintentional replies, and the functionality operated as expected. The presentation gave little information about how OpenAI will handle security related to facial recognition and speech creation, despite Murati’s declaration that the business was taking precautions to stop its new voice capabilities from being abused. Bloomberg claims that OpenAI is also nearing completion on an agreement to integrate its generative AI capabilities into Apple’s iPhone operating system. Since it does not yet have a generative AI product that is available to the public, similar to competitors like Google’s popular Gemini model, Apple has been making an effort to redirect resources towards AI. It seems to have more sophisticated functions in its voice mode than Apple’s Siri voice assistant can give.
In addition to growing its product line and pursuing partnerships, OpenAI is being sued by media companies and publishers for alleged copyright infringement. News outlets including the New York Times are suing OpenAI, claiming it is illegally using their work to train its AI models without permission or payment, and are pursuing damages that might total billions of dollars.
Leave a Reply