New ChatGPT Version Unveiled with 'Real Voice Conversation' Feature

OpenAI, a leading American artificial intelligence development company and creator of ChatGPT, has unveiled a new AI model named 'GPT-4 O'. This model, which is an improved version of the GPT-4 language model, was introduced on Monday, 24 Ordibehesht, at a special event held by the company. It is capable of real voice conversation and mixed interaction between text and images. The letter O (0) that follows the number 4 stands for Omni, a Latin prefix meaning 'all', and OpenAI has used it to signify the global and universal nature of this product. The new feature allows users to speak with ChatGPT and receive instant, real-time responses, and it is now also possible to interrupt ongoing conversations with ChatGPT. According to OpenAI experts, these two features are both hallmarks of real conversations that have not been present in AI voice chatbots until now. Mira Moriti, Chief Technology Officer of OpenAI, announced that the GPT-4 O capability will be offered for free due to its greater efficiency compared to the company's previous models. However, users who purchase this version will have access to more features compared to free users. During the unveiling of the new feature, OpenAI researchers showcased the capabilities of the new ChatGPT voice assistant. In one example, the voice assistant was able to narrate a bedtime story with various sounds, emotions, and tones. In another example, the ChatGPT voice assistant used its visual capabilities to solve a mathematical equation written on paper step by step. The simultaneous and instant translation capability of GPT-4 O was also demonstrated. Experts say the next step for ChatGPT is to achieve a search capability similar to search engines for accessing and linking to up-to-date and accurate web information, a feature that the current version of ChatGPT struggles with. OpenAI has announced that it will release this new AI capability in the coming weeks.