Even when using ChatGPT with the permission of your professor, and with citing ChatGPT as a source, there are some ethical conundrums posed by the use of ChatGPT in academia. Some common concerns about the use of ChatGPT in college are:
- Not always factual - ChatGPT does not fact check the information that it uses to create its language databases. Thus, the answers provided by some queries to ChatGPT may be incorrect (University of Michigan Library, 2023).
- Incorrect or made-up citations - When asked to provide citations, ChatGPT does not always provide accurate, or even real, citations. Often, when its own database is insufficient, ChatGPT will “hallucinate” or make-up information that it anticipates its end-users require (Scheelke, 2023).
- Currency - ChatGPT was provided with training material up until September of 2021. This means that its databases don’t include current, up-to-date information on timely topics and issues (Scheelke, 2023).
- Racism and biases – ChatGPT can be manipulated into providing racist responses (Alba, 2022). In addition, the technology was created using content made primarily by white men of privilege, meaning its answers may be biased or skewed towards favouring current hierarchies in society (Farmer & Croxton, 2023).
- Labour issues – ChatGPT used workers in Kenya, which they paid less than $2 an hour, to train its AI to recognize examples of violence, hate speech, and sexual abuse. Some workers failed to receive proper counselling supports (Perrigo, 2023).
- User privacy – ChatGPT requires users to create an account to use the service and collects information from users for its own purposes. It’s a good idea to familiarize yourself with ChatGPT’s privacy policy before using the service (Centennial College Libraries, 2023).
- Copyright – ChatGPT uses information derived from the internet without necessarily offering the original author(s) of that information with either credit or compensation (Centennial College Libraries, 2023 & Valdez, 2023).
