Let's train an AI to mirror a specific style, such as the work of Mahatma Ghandi, or your chat history. We will gain experience collecting training data (which may be stored in hard-to-access forms such as chat logs or PDF's), training a machine model using this data, and potentially even deploying it somewhere. Along the way, we will learn how to use language models to interrogate the data that we trained it on. 

An example is this project I worked on, where we trained a language model to write like an Instagram influencer: https://www.instagram.com/myfriendsylvia/?hl=en

Build an AI clone

Journalists and social scientists are turning to NLP to understand complex relationships among documents and the entities they reference. An example is a story I worked on for WIRED, for which we built a graph database out of Twitter posts, and used this to discover mistakes in Twitter's disinformation policy. 

https://www.wired.com/story/how-americans-wound-up-on-twitters-list-of-russian-bots/