UD Logo
School of Education

ChatGPT & AI Chatbots

Module 1. What Are AI Chatbots?

AI chatbots are intelligent conversational agents powered by large language models pre-trained on massive databases (such as the entire Wikipedia or the Common Crawl) that are used in generating responses by transforming what the user types. The company OpenAI calls this process a Generative Pre-trained Transformer (GPT). The first version of GPT appeared in 2018. Subsequent versions are numbered as in GPT-2, GPT-3, and GPT 3.5, which powers ChatGPT. A subsequent version called GPT-4 accepts both text and images as input. Microsoft invests heavily in ChatGPT, which Microsoft plans to integrate across its product lines including Bing search, Office 365, GitHub, and Visual Studio.

At the same time as OpenAI was working on ChatGPT, Google was developing an AI Chatbot called Bard. When Microsoft released ChatGPT and Google saw how it went viral by amassing more than a million users in its first week, Google felt some pressure to release its own chatbot. Later in this course you will come to a module about Bard, which is powered by Google’s Pathway Language Model (PaLM).

More AI is coming. Amazon is working on Bedrock. Meta Facebook has a Large Language Model called LLaMa. OpenAI has a DALL-E deep learning model that can generate AI images in response to natural language descriptions. Here are some recommendations from ZDNet about which tools work best: