Understanding the generative AI landscape

Since the advent of ChatGPT, organizations across the globe have explored a plethora of use cases that generative AI can solve for them. They have built several innovation teams and teams of data scientists to build and explore various use cases, including summarizing long documents, extracting information from documents, and performing sentiment analysis to gauge satisfaction or discontent toward a product or service. If you have been working in the machine learning (ML) or natural language processing (NLP) field, you may be familiar with how a language model works – by understanding the relationship between the words in documents. The main objective of these language models is to predict the next probable word in a sentence.

If you look at the sentence John loves to eat, a natural language model is trying to predict what the next word or token in the sequence will be. Here, the next probable word seems to be ice-cream, with a 9...