Student Guide to Generative AI

What is Generative Artificial Intelligence (AI)? 

Generative AI is a complex technology that powers software and tools which can be asked questions in conversational language and produces replies that read as if they were written by a human being. The technology can also learn and improve based on more data and inputs that it encounters. 

How does Generative AI work?

Generative AI is a type of artificial intelligence (AI) that uses machine learning algorithms to create new and original content like images, videos, text, and audio. 

• The AI tool is made up with a massive repository of information as well as a neural network which is uses to "think" to create text or media files like images, text, data, or sounds, depending on the tool. 
• The user provides the AI with a request or sample of the desired content. 
• The AI uses its neural network to generate responses that uses the ones it has trained from to help "predict" the best response that the use will want. 

How does Generative AI actually create responses?

It creates its answers by analyzing enormous amounts of existing text, which enables it to mimic sentence structure and argument, and "predicts" the best answer. When asked a question or presented with a statement, a chat AI finds information associated with the terms it's given and generates text about it, using algorithms that determine the next word based on its analysis of millions of existing, human-generated sentences. The mechanism is similar to the predictive text feature on a cell phone, though vastly more sophisticated. The result is usually clear, grammatically correct sentences.

(Image Credit:This work is licensed under the Creative Commons Attribution - Noncommercial - ShareAlike 4.0 International License. Created by Jessica Kiebler, Pace University.)

Where does the information come from?

The algorithm is trained on massive corpus of text data, around 570GB of datasets, including web pages, books, and other sources in order to compile its answers. This is why the underlying technology is called a large language model or LLM as they are built by "learning" patterns in large amounts of information.  

When asked a question or presented with a statement, a chat AI finds information associated with the terms it was given and generates text about it, using algorithms that determine the next word based on its analysis of millions of existing, human-generated sentences. The mechanism is similar to the predictive text feature on a cell phone, though vastly more sophisticated. The result is grammatically correct, reasonably well-organized text that reads uncannily like it was written by a person.

What is ChatGPT?

An example of generative AI is ChatGPT which was created by the company OpenAI and released in late 2022 to the public. It can produce an essay, an explanation of a natural phenomenon, a meal plan, a vacation suggestion, or answers to a vast number of other questions.

Are answers limited to text?

ChatGPT's output is not limited to ordinary prose. It can write code, create an outline, even craft a poem or song. It can be asked to write or rewrite in a particular style, or to revise and clean up written text. However, if the AI does not have access to accurate information, it will still generate text -- it just won't be accurate.

There are many other types of generative AI which can create images and artwork as well.