Topic 11g
ChatGPT

ChatGPT is an AI that uses a Large Language Model (LLM) which employs unsupervised learning and neural networks while raising a significant number of ethical issues. We are including it in this course not because we feel that you need to have deep understanding of how it works or what it does. However, we feel that a basic understanding and discussion about how it works can serve as a health wrap-up to this chapter.

Learning Outcomes

By the end of this topic students should be able to:

• Explain how ChatGPT works (in very general terms)
• Evaluate the strengths and limitations of ChatGPT as a language model, including its ability to generate coherent and relevant responses, and its potential for bias or error.
• Discuss how a tool such as ChatGPT could play a future role in domains such as customer service, language translation, and content generation, and be able to identify ethical and societal implications of such use.

• ChatGPT passes the Turing Test
  • https://www.techradar.com/opinion/chatgpt-has-passed-the-turing-test-and-if-youre-freaked-out-youre-not-alone
• Large Language Models
  • https://blogs.nvidia.com/blog/2023/01/26/what-are-large-language-models-used-for/
  • https://blog.dataiku.com/large-language-model-chatgpt
• Why ChatGPT isn't going to completely replace programmers
  • https://futurism.com/the-byte/study-chatgpt-answers-wrong

In class

Discussion

Let's walk through this idea for a bit.

Checking for Understanding

• Some people feel like we should fight tools like ChatGPT vigorously while others claim that, instead, we should understand where the use of tools such as these are helpful and what skills we can teach/learn to better work with them. What discussion points in this argument seem "valid" considerations to you?
• This kind of argument isn't actually that new. These same people probably fought about the use of graphing calculators (vs teaching kids to graph with paper/pencil) or the use of word processors with spell checkers (vs typewriters). In what way(s) are these historical arguments related to this current argument. In what way(s) might we claim that this argument is very different?

Again, I don't think there are right/wrong answers here. This is all about the discussion.