Following current usage, this document uses generative AI to refer to services that generate text, images, or other media using large language models or related technology. Such services include Microsoft Copilot, ChatGPT, Gemini 1, and many others. The use of generative AI in my courses will be governed by the policy in this document. Specific deliverables may have additional restrictions.
The term "someone else" in the UNB plagiarism policy will be interpreted as the creator of the original work 2 used to build the model backing a generative AI service. This creator is almost certainly not the provider of the generative AI service 3.
To put the previous point another way, if you use the output from a generative AI service in your coursework, you are responsible for citing the original source, not only the generative AI service. The penalties for failing to do so, including for failing to verify a source claimed by a generative AI service, are defined by the UNB plagiarism policy.
Generative AI services are not acceptable as sources of facts. If you use them to locate sources of facts, you are responsible for verifying that those sources actually exist, and say what the generative AI service claims.
You may use automated tools to provide suggestions and writing feedback, but copying and substantial blocks of text from a generative AI service will be considered as be considered either as plagiarism or as non-completion of the relevant deliverable, depending on whether the strict crediting rules above are satisfied 4.
- These names are trademarks of Microsoft, OpenAI, Google (Alphabet) respectively.↩
- In the machine learning literature, and some popular writing, this is referred to as training data.↩
- Yes, this means (the owners of) most generative AI services are guilty of plagiarism, but they are not enrolled in my course.↩
- Essentially, if you can't be bothered to write it, we don't want to read it.↩