How Does Generative AI Work?

By Ojas Joshi

ChatGPT, Bing, Google Bard—recently, the world has been taken by storm by the introduction of a range of open-access language-based artificial intelligence models. Revolutionary, chatbots such as ChatGPT utilize a breakthrough development in A.I. technology: generative A.I. Through generative A.I., models have the ability to teach themselves to predict and spot patterns in text, music, and art. Harnessing multiple terabytes (1 million data bits) of information from many sources (books, articles, plays, scientific journals), generative A.I. models use their ability to predict words and patterns to build sentences, essays and plays. Moreover, generated information grows more powerful over time as the model adapts in response to user input. Generative A.I. models can process scores of unstructured data, making them adaptable to any situation. They are not designed to perform a single function; rather, they use data to meet the users needs. Furthermore, the likes of ChatGPT and Google Bard use unsupervised learning (learning without human involvement) to efficiently store and use data.