Artificial Intelligence (AI) and Information Literacy

You might have heard about or used artificial intelligence-based tools like ChatGPT or DALL-E, but how do these tools actually work? This section will cover how tools like these are made, how they generate information, and some considerations to think about as you assess if and how to use them.

What is Generative Artificial Intelligence?

A lot of the internet uses some form of AI to run. For example, a type of AI called discriminative (or predictive) AI creates all those shopping recommendations on Google and algorithms on TikTok.

In general, AI just means a set of rules used by computers to try and act like what we think of as human intelligence.

Generative AI is a specific type that tries to create more information that looks like the information it has been trained on, following the same patterns. For example, if it has looked at a bunch of cats, it will attempt to spit out an image that has four legs, a tail, and pointy ears that hopefully looks like a cat.

Generative AI often “hallucinates” and creates facts or sources that don’t exist. It doesn’t know how to tell what’s real or not because it’s not actually thinking.

Generative AI is not a search engine and is not searching for information. Instead, AI is creating new content in response to what you say.

Be sure to fact check anything you get from AI.

What is actually going on when you use an AI-based tool? Check out this three-minute explanation from Hal Daumé III, Professor of Computer Science at University of Maryland, Institute Director for the Institute for Trustworthy AI in Law & Society (TRAILS).

Here are four facts about ChatGPT that will be helpful as you think about its uses.

1. ChatGPT is developed by a company called OpenAI, based in San Francisco, CA. Their stated mission is "to ensure that artificial general intelligence—AI systems that are generally smarter than humans—benefits all of humanity." Chat GPT is just one of OpenAI's products, which also include DALL-E, the text-to-image generator. OpenAI begun as a nonprofit in 2015 then became a for-profit company in 2019. They received investments from different sources, including $1 billion from Microsoft, which uses the same AI model that powers ChatGPT and DALL-E for Bing AI.  
2. It can generate new text for any written prompt and is designed to provide human-like conversational interactions. Some examples of things it can do include: write an essay, generate a definition, debug or write functional code, suggest tourist destinations, make up recipes based on a collection of ingredients, craft a love letter, etc.
3. It produces unique responses each time, and you can tweak its responses because it remembers the whole chat conversation. You can instruct it to respond to you in different ways (e.g. "please reply to me like I'm in 2nd grade").
4. The "GPT" in ChatGPT stands for Generative Pre-trained Transformer. 

   "Generative" refers to a type of machine learning model that creates new content.

"Pre-trained" refers to a type of machine learning model that was trained on an enormous amount of data. In the case of ChatGPT, the model was trained and on hundreds of billions of words (mostly from websites) to learn patterns and relationships between words and phrases.

"Transformer" refers to part of the machine learning model that can better understand sentences because it can understand the relative importance of different words in context.

ChatGPT-3.5 and ChatGPT-4 don't have the ability to search the Internet in real time (although OpenAI and other companies are working on that!). We don't know for sure what data OpenAI used to train ChatGPT, but many AI programmers train large language models on Google's C4 dataset. Check out this Washington Post article on what websites are included in that dataset.