Skip to Main Content

Faculty Help: Generative AI Resource Guide: ChatGPT 101

Repository of info about impact of Generative AI on/in higher education.

The "101" on ChatGPT

For this section I will let others more knowledgeable than myself answer these questions. 🙂
Click a question below to be taken to the answer. 

I'd like to sign up and experiment with ChatGPT. How do I do that?

What is ChatGPT? 

What's the difference between the various versions (3.5, 4.0 or 4o) of ChatGPT? 

How exactly does ChatGPT work? (jargon-free, please!)

How exactly do LLMs work? (jargon-free, please!)

Is ChatGPT's output reliable? Can I trust the veracity of its information?

Have a lot of people heard of ChatGPT? How many people are aware of and using it?

*sigh* Is this just another hyped up tech phenomenon?

Does the ChatGPT site have its own FAQ section?

I hear OpenAI employed Kenyan workers to train ChatGPT to filter out harmful content, including violent and sexual material. Is this true?

What are some other ethical implications I should consider when engaging with ChatGPT?

 

I'd like to sign up and experiment with ChatGPT. How do I do that?

ChatGPT: https://chat.openai.com/

ChatGPT Overview: https://openai.com/chatgpt/ 

 

What is ChatGPT?

 

What's the difference between the various versions (3.5, 4.0 or 4o) of ChatGPT?

Both 3.5 and 4o are free. However, 4.0 currently costs $20/month.

From OpenAI's website
GPT-4o is our most advanced multimodal model that’s faster and cheaper than GPT-4 Turbo with stronger vision capabilities. The model has 128K context and an October 2023 knowledge cutoff. (More info on version 4o here.)

GPT-4 Turbo is offered at 128K context with an April 2023 knowledge cutoff and basic support for vision.

GPT-3.5 Turbo is our fast and inexpensive model for simpler tasks. gpt-3.5-turbo-0125 is the flagship model of this family, supports a 16K context window and is optimized for dialog. gpt-3.5-turbo-instruct is an Instruct model and only supports a 4K context window.

 

How exactly does ChatGPT work? (jargon-free, please!)

This visual, interactive guide from The Guardian does an excellent job of explaining!

How AI chatbots like ChatGPT or Bard work – visual explainer | The Guardian

 

How exactly do LLMs work? (jargon-free, please!)

This is one of THE best explanations we've seen on the interwebz thus far. Enjoy!
A jargon-free explanation of how AI large language models work | ArsTechnica

And here are some handy glossaries of ChatGPT terms:

Artificial Intelligence (AI) Terms: A to Z Glossary | Coursera

Basic Terms All ChatGPT Users Should Know | CNET

ChatGPT glossary | Geeky Gadgets 

A ChatGPT Glossary | WIT Lingo

 

Is ChatGPT's output reliable? Can I trust the veracity of its information?

That's not recommended right now. Because of how LLMs work and are trained its output can indeed sound very plausible and authoritative. However, a phenomenon known as hallucination is still common. Additionally, at this time (late spring 2024) ChatGPT is not capable of fact-checking its output and it has no true understanding of source provenance. In conclusion, if what you are researching matters (and it always should! 🙂) then it is recommended you fact check and verify ChatGPT's output.

From the website: "ChatGPT can make mistakes. Check important info."      

 

Have a lot of people heard of ChatGPT? How many people are aware of and using it?

About 1 in 5 U.S. teens who’ve heard of ChatGPT have used it for schoolwork | Pew Research | Nov.16, 2023

Most Americans haven’t used ChatGPT; few think it will have a major impact on their job | Pew Research | Aug.28, 2023

 

 

*sigh* Is this just another hyped up tech phenomenon?

Probably not - it's likely here to stay. But have a look at the Gartner Hype Cycle chart to better understand how these things often evolve over time. Learn more about the Gartner Hype Cycle here and here 

 

Does the ChatGPT site have its own FAQ section?

Yes. Here it is https://help.openai.com/en/articles/6783457-what-is-chatgpt

And here's a helpful infographic to get you started. 

 

I hear OpenAI employed Kenyan workers to train ChatGPT to filter out harmful content, including violent and sexual material. Is this true?

Yes, this is true. 

Reference & Instruction Librarian

Profile Photo
Sarah Hood
Contact:
505.428.1830
chat loading...