Skip to Main Content
Skip to Library Help widget
 

ChatGPT

What is ChatGPT?

ChatGPT is a natural language processing tool driven by Artificial Intelligence (AI) technology that allows users to have human-like conversations and much more with the chatbot. The language model can answer questions and assist with tasks such as composing emails, essays, and code.

The dialogue format makes it possible for ChatGPT to answer followup questions, admit its mistakes, challenge incorrect premises, and reject inappropriate requests.

Usage is currently open to the public free of charge because ChatGPT is in its research and feedback-collection phase. A paid subscription version called ChatGPT Plus was launched February 2023.

ChatGPT was created by OpenAI, an AI and research company. The company launched ChatGPT on November 30, 2022. 

What can ChatGPT do and not do?

ChatGPT can: 

  • Write an essay, an email, song lyrics, a poem
  • Create an app
  • Write code
  • Build a resume
  • Write Excel formulas
  • Summarize content
  • Write a cover letter
  • Compile reading or research lists
  • Generate ideas or brainstorm

ChatGPT can't:

  • Write a self-reflection
  • Write about anything that happened after 2021
  • Provide non-text based responses
  • Make predictions about future events
  • Browse or summarize content from the internet
  • Draw connections between class content and visual materials

While ChatGPT can be useful for some tasks, it does have limitations, as noted by its creator OpenAI: 

  • ChatGPT sometimes writes plausible-sounding but incorrect or nonsensical answers. The ideal answer depends only on what the model knows, rather than evaluate the veracity or accuracy of that knowledge. Answers may include biased or misleading information. 
  • ChatGPT is sensitive to tweaks to the input phrasing or attempting the same prompt multiple times. For example, given one phrasing of a question, the model can claim to not know the answer, but given a slight rephrase, can answer correctly.
  • The model is often excessively verbose and overuses certain phrases, such as restating that it’s a language model trained by OpenAI. 
  • Ideally, the model would ask clarifying questions when the user provided an ambiguous query. Instead, current models usually guess what the user intended.
  • While efforts have been made to make the model refuse inappropriate requests, it will sometimes respond to harmful instructions or exhibit biased behavior. Certain types of unsafe content is blocked, but there are some false negatives and positives. 

Additional limitations: 

  • "Hallucination." ChatGPT will fabricate information or provide false information in order to fill gaps in its knowledge. The model cannot effectively distinguish between accurate, factual information and false information. Read more about hallucination in Hacker News.
  • ChatGPT cannot provide information after 2021, so questions about current events may be inaccurate or return no results; it also cannot provide information pre-internet.  
  • ChatGPT does not provide sources for the information in its responses. However, there is a workaround to view the sources used by ChatGPT in its response.

Sources

Gerwitz, David. "How to make ChatGPT provide sources and citations." ZDnet.com. Apr. 14, 2023. https://www.zdnet.com/article/how-to-make-chatgpt-provide-sources-and-citations/

msravi. "ChatGPT produces made-up nonexistent references." Hacker News. https://news.ycombinator.com/item?id=33841672

OpenAI. "Introducing ChatGPT." https://openai.com/blog/chatgpt. 

Report ADA Problems with Library Services and Resources. Resources on this page may require Document Viewers