@@ -86,24 +86,29 @@ To facilitate this, and to be very clear
# Chat GPT
It is hard to find someone who has not heard about ChatGPT and related tools, and these tools are undeniably useful for generating ideas, providing suggestions, and more. However, we recommend that you use generative AI in moderation. Generative AI can help you to summarize text, improve grammar, write code, collect relevant resources to read, and generate ideas. It can also [lead you to misinformation](https://medium.com/@emilymenonbender/cleaning-up-a-baby-peacock-sullied-by-a-non-information-spill-d2e2aa642134) and is generally less powerful than the hype might lead you to believe, as eloquently described by UW faculty member Emily Bender in her [Seattle Times OpEd](https://www.seattletimes.com/opinion/look-behind-the-curtain-dont-be-dazzled-by-claims-of-artificial-intelligence/). If you are interested in learning more about generative AI's limitations, here is a [helpful repository](https://faculty.washington.edu/ebender/media/).
It is hard to find someone who has not heard about ChatGPT and related tools, and these tools are undeniably useful for generating ideas, providing suggestions, and more. However, we recommend that you use generative AI in moderation. Generative AI can help you to summarize text, improve grammar, write code, collect relevant resources to read, and generate ideas. It can also [lead you to misinformation](https://medium.com/@emilymenonbender/cleaning-up-a-baby-peacock-sullied-by-a-non-information-spill-d2e2aa642134) and is generally less powerful than the hype might lead you to believe, as eloquently described by UW faculty member Emily Bender in her [Seattle Times OpEd](https://www.seattletimes.com/opinion/look-behind-the-curtain-dont-be-dazzled-by-claims-of-artificial-intelligence/). AI can be **biased** ([including disability bias](https://www.youtube.com/watch?v=XFgGEfWf3UY)), **flat out wrong**, is built on writing by other people that it **might plagiarize** and probably doesn't know much about this course or how to do well in it. If you are interested in learning more about generative AI's limitations, here is a [helpful repository](https://faculty.washington.edu/ebender/media/).
In addition to the abovementioned concerns, if you start to rely solely on generative AI, you may limit your own development in critical thinking and writing, and if the results are that your writing is narrower and shallower in scope this may impact your grade.
In this class, we will ask you to follow these ethical guidelines when using generative AI such as ChatGPT.
In this class, we will ask you to follow these ethical guidelines when using generative AI such as ChatGPT.**NOTE** this policy most likely will evolve over the quarter, since this is such a new situation. I will announce modifications and expect you to abide by them.
1. You may use generative AI for assignments and exams unless I specify that it may not be used. If guidelines are provided you should additionally follow those guidelines.
2. Unlike blog posts and research articles, you do not need to attribute artifacts/quote text or label code produced by generative AI when you use it. However, you must do the following or you will face academic consequences including but not limited to failing an assignment or an exam.
- You must cite the AI program you used in the artifact you hand in
- If it copies text from other sources and you don’t put that text in quotes and provide proper attribution, you will be held accountable for that.
- If it provides ideas closely based on other sources and you don't cite those sources, you will be held accountable for that.
- You must still comply with the academic integrity policies of the institution. This includes refraining from using generative AI to plagiarize or cheat.
- You must still comply with the academic integrity policies of the institution. This includes refraining from using generative AI to plagiarize or cheat. For example, you may not use AI to create fake data (accessibility reviews, interview data, user data, etc.) and pass that off as based on real people.
3. You will be held to the same standards when you use generative AI as for any assignment, regardless of whether you or the AI created something, including:
- If you turn in artifacts that are not accessible, you will be graded accordingly
- If you turn in artifacts that contain false or incomplete claims, you will be graded accordingly
- If you turn in code that does not compile or is incomplete, you will be graded accordingly
- You will be graded based on the critical thinking and writing skills, accuracy, and accessibility of the things that you produce.
Breaking the rules above, or generally using AI tools in ways that
are nefarious or unacknowledged, is academic misconduct and may be
subject to the misconduct policies detailed here.
Also please note that using such tools has consequences unrelated to your schoolwork. For example, generative AI models use an [enormous amount of water](https://arxiv.org/pdf/2304.03271.pdf) and [energy](https://www.nature.com/articles/d41586-023-00843-2): ``ChatGPT needs to “drink” a 500ml bottle of water for a simple conversation of roughly
20-50 questions and answers, depending on when and where ChatGPT is deployed.''
In addition, you are probably donating your data to the companies that deployed them. Please take reasonable steps to avoid making our assignments easier in future iterations of the course (e.g., once the tool provides a correct answer, don't give it positive feedback).
