Flowchart devised by Aleksandr Tiulkanov, AI and Data Policy Lawyer, January 2023, retrieved from UNESCO's ChatGPT and Artificial Intelligence in higher education: Quick start guide. Please note that this flowchart isn't comprehensive regarding the risks of using ChatGPT; see What are some of the risks of using ChatGPT? on this page for more information.
In response to the prompt of "Succintly describe some of the the risks of using ChatGPT", the tool provided the following response. Note that the response isn't an exhaustive list of risks; all responsibility for using the tool rests with users:
Using ChatGPT comes with certain risks that should be considered. These include:
Inaccurate or misleading information: ChatGPT generates responses based on patterns in its training data, which means it may occasionally provide incorrect or unreliable information. It's important to verify the accuracy of the responses and not solely rely on them.
Lack of ethical judgment: ChatGPT does not possess moral or ethical values. It may generate responses that are inappropriate, biased, or offensive. Users should be cautious and apply their own judgment when interacting with the model.
Amplification of existing biases: If the training data contains biases or stereotypes, ChatGPT may inadvertently reflect or amplify them in its responses. This can perpetuate societal biases and create a distorted view of certain topics.
Privacy and security concerns: Interactions with ChatGPT can involve sharing personal or sensitive information. Users should be cautious about sharing such details and ensure they are using the model in a secure and trustworthy environment.
Unintended consequences: Due to the complexity of language and potential limitations of the model, there is a possibility of ChatGPT generating unexpected or unintended responses. These responses may have unintended consequences or implications.
Understanding these risks and exercising critical thinking when using ChatGPT can help mitigate potential issues and ensure a responsible and informed interaction with the model. (OpenAI, 2023)
For a more information on the risks of using ChatGPT, please see "Challenges and Ethical Implications" in UNESCO's (2023) ChatGPT and Artificial Intelligence in Higher Education: Quick Start Guide (pp. 10-11).
Image by Gerd Altmann from Pixabay
Reference
OpenAI. (2023). ChatGPT (May 9 version) [Large language model]. https://chat.openai.com/chat
Students working with ChatGPT should be cautious about the information they're providing when using the tool.There are broad implications for students using ChatGPT who are sharing their intellectual property with the tool to generate text e.g., sharing original research results, especially if doing so compromises the privacy of research subjects.There are also considerations around sharing protected information, such as proprietary information, and culturally sensitive information, such as Indigenous Teachings. Once the tool has been trained on that information, it is impossible to remove the details from the tool. For more information, see "Challenges and Ethical Implications" on pages 10-11 in ChatGPT and Artificial Intelligence in Higher Education: Quick Start Guide.