OpenAI’s most recent version of its artificial intelligence model can emulate human cadences in spoken responses and even attempt to discern people’s emotions.
The effect is reminiscent of Spike Jonze’s 2013 film “Her,” in which the (human) main character falls in love with an artificially intelligent operating system, resulting in issues. While few will find the new model appealing, OpenAI claims it is faster than earlier iterations and can reason over text, audio, and video in real time.
GPT-4o, which stands for “omni,” will power OpenAI’s popular ChatGPT chatbot and will be available to everyone, including those using the free version, in the coming weeks, the company revealed during a brief live-streamed update.
CEO Sam Altman, who was not a presenter at the event, simply put the word “her” on the social networking platform X.
During a presentation with Chief Technology Officer Mira Murati and other executives, the AI bot interacted in real time, adding emotion — especially “more drama” — to its voice when asked.
It also assisted in walking through the steps required to solve a simple math issue without initially spitting out the answer, as well as a more difficult software coding challenge on a computer screen.
It also attempted to extrapolate a person’s emotional state by looking at a selfie video of their face (deciding he was pleased because he was smiling) and translated English and Italian to explain how it may help people with different dialects have a conversation.
Google plans to have its I/O developer conference on Tuesday and Wednesday, during which it is likely to reveal upgrades to its own Gemini AI model.
Also Read: Microsoft Plans to Invest $1.7 Billion in AI and Cloud Computing in Indonesia
