THE EVOLUTION OF CHATGPT: FROM TEXT TO VOICE IN VIRTUAL WORDS

Book excerpt: ChatGPT is an artificial intelligence (AI)-based chatbot that makes use of natural language processing in order to provide conversational discourse that sounds and feels like it was done with a real person. This is accomplished by producing conversational discourse in many languages, including English, Japanese, and Korean. ChatGPT is a cloud-based application that may be accessed using the Azure cloud platform. It was created by Microsoft Research. The language model is able to provide responses to inquiries as well as generate a variety of written material, such as emails, posts on social networking sites, and posts on platforms that enable users to write in the format of an essay. On November 30, 2022, OpenAI developed ChatGPT, a big language model-based chatbot, and made it available to the public. Only the beta version of ChatGPT was first made accessible to users. Through the utilization of this feature, users are provided with the capacity to enhance and lead a discussion in the direction of a desired duration, structure, style, level of information, and language. This ability is made available to users. What we refer to as the Chat Generative Pre-trained Transformer is referred to as ChatGPT in its shorter version. In order to determine the context of the dialogue at each level of the discussion, a method known as "prompt engineering" is utilized. This method involves looking at the prompts and responses that came before it in the conversation. ChatGPT is built on either the GPT-3.5 or the GPT-4 model, both of which are a part of OpenAI's proprietary line of generative pre-trained transformer (GPT) models. Both models were used in the construction of ChatGPT. Conversational interactions are one of the primary focuses of ChatGPT's design. The Google-developed transformer architecture serves as the foundation for these models, which have been modified for use in conversational applications by employing a mix of supervised learning and reinforcement learning strategies. These methods were utilized in order to get a higher level of precision with the models. The research preview version of ChatGPT was made accessible to the general public for the very first time as a free service.