Gábor’s AI policy for advanced applied metrics classes and projects

  • V.0.2.6. 08. February 2024.
  • This is a policy is written for my advanced classes: Applied Economics with Sports Data, Data Analysis 3 Prediction and ML. Data Analysis 4 Causal Analysis. It also applies to term papers and theses.
  • Gábor Békés (Central European University, BekesG@ceu.edu)
  • Licence: CC BY-NC 4.0 free to share and adapt.

Introduction: Enjoy “AI” with moderation

  1. The abundance of Large Language Models (LLM) driven “AI” means that the process of doing data analysis projects has changed, and these tools are becoming essential help. I define “AI” here as LLM and other machine learning methods including ChatGPT, Github Copilot, Bard/Gemini, Claude, Grammarly, DeepL and any other available tools. Prompting means communication with “AI” tools such as asking questions or checking code.
  2. This policy is aimed at more advanced classes and assume that you already have core coding and statistics/econometrics skills. In this course, you will be advised to use new tools but at the same time remain able to do most work without them: enjoy “AI” with moderation.

Suggestions and advice

  1. All evidence suggests you shall use the most developed versions. Currently this is ChatGPT4. Google’s Gemini Ultra is also promising. Using a lower version such as Bard/Germini and ChatGPT 3.5 means more hallucinations and an overall less adequate experience for professional / academic use.
  2. Consider experimenting with AI based services helping with literature search and review. Currently one the most widely used one is Consensus It’s still beta, but has a nice GPT. My experience is mixed.
  3. Consider experimenting with Github copilot including as add-in in VScode, RStudio, etc. It is helpful in speeding up coding, and even writing code based on prompts. Extensive and deep debugging is needed, as code is likely to contain bugs. Github copilot has an education offer.

Policies for assignments, papers, presentations

  1. You are responsible for all aspects of your work. AI is like search, it’s helpful but use it as an input. You must carefully check and debug any input from AI.
  2. You may use any “AI” assistance for writing your report or presentation. You may build on output following your prompts, but consider any output material by “AI” as an input to your project
  3. Any input from AI shall be treated similarly to any other source such as a book or a paper: you must reference it by linking the public link generated for the conversation. Not doing so will be considered a serious ethical violation equivalent to plagiarism.
  4. In order to ensure the best learning outcome, I suggest using “AI” for the following tasks only
  5. Drafting code snippets, turning code bits into functions,
  6. Improving code, commenting it
  7. Idea generation, thinking about predictors, causal pathways, potential confounders
  8. Potential problem finding, discussions on what can go wrong
  9. Language improvement, translation
  10. Technical help re markdown/latex/beamer/ppt, formatting
  11. Learn more about topics by having a conversation with AI on the material. You may copy text into it and ask for help in understanding details, and clarifications.

Policies for coding

  1. While I expect you to be able to code without “AI”, in your assignments, you may use any “AI” assistance for coding. Note that you carry all responsibility for the correctness of your code. Please test and debug your code relentlessly.

Policies for exam, quiz

  1. All quiz, exams will be on site with zero “AI” tool use policy. Any use of “AI” tools will be considered cheating. This may require provision of offline computers or pen and paper exams.

Background assumptions

This policy is based on a few assumptions

  1. Using “AI” tools to assist your work is a productivity enhancer
  2. The ability to use “AI” as copilot, carry out prompt engineering is a new and important skill to learn and practice.
  3. There is no way to police the use of “AI” tools, so no point in trying. This implies ethical consideration for students – they shall be encouraged to show self-restraint for the benefit of their future self.
  4. As a consequence of the no policing argument, It is no longer possible to have take-home exams.
  5. It is essential for students to find a balance between learning the material as well as get practice with “AI” tools.

Fine print

  1. When link to chat is not natively supported, you may use third party plugins. If neither is supported please use a link supporting service (e.g Chatgpt) instead.