Skip to Main Content

ChatGPT and other Generative AI tools

What is Generative AI?

Generative AI is a type of artificial intelligence that creates new and original content, like art, music, or text, by learning patterns from existing data. It uses complex algorithms and models to generate content that resembles human-created examples. This technology finds applications in art, content creation, data science, fashion, chatbots, and more. However, it's important to remember that generative AI doesn't truly understand or think—it replicates patterns it has learned.

What is ChatGPT?

ChatGPT is probably the most popular generative AI tool at this time. ChatGPT is a chatbot that produces human-like text responses to written prompts. While ChatGPT provides very human-like responses, the information it provides may not always be accurate or reliable, so it's a good idea to verify important information from trusted sources. 

Where does the information in ChatGPT come from?

ChatGPT is trained on vast and varied data from the internet including books, articles, websites, and other written content. While some of the contents of the dataset it is trained on is publicly available, the full extent of the dataset is unknown. 

What we DO know is that the free version of ChatGPT is trained on data from April 2023 and earlier, which means when looking for text with information on current events, ChatGPT will not provide accurate information. The paid version of ChatGPT is trained on data through June 2024, but can fetch current information from the internet.

Issues with ChatGPT

  • Inaccurate and biased information: ChatGPT is trained on a vast amount of data that may or may not be accurate, and thus may provide responses with inaccurate or even biased information. While ChatGPT does not provide sources or references, it may do so if asked. However, the sources it provides may also be inaccurate or nonexistent. ChatGPT may also invent/hallucinate information.
  • Privacy: Using ChatGPT requires an account with OpenAI which allows the input of personal information without the knowledge of how that information will be used.
  • Currency: The unpaid version of ChatGPT is trained on internet data only through September 2021, so it cannot produce accurate responses about current or realtime information
  • Plagiarism: The use of ChatGPT or other Generative AI tools is seen as plagiarism and academic dishonesty if used in an assignment without explicit permission of your teacher.
  • Labor Exploitation: There have been reports of the outsourcing and exploitation of labor to train data on AI tools such as ChatGPT.

Critically evaluating information in ChatGPT (and other AI) responses

Because ChatGPT and other generative AI tools are trained on data that may be biased, outdated, or incorrect, it is imperative that we critically evaluate the information contained in their outputs. Using these strategies will help you evaluate the information:

  • Lateral Reading: look for other trusted sources that can verify the information provided by the AI
  • Verify Sources: confirm any sources that are provided by AI tools exist (AI tools are known to create false information, including sources)