As more library databases incorporate Large Language Models (LLM), generative artificial intelligence (AI), and other emerging technologies, your students may question how using the library tools they know may interact with your syllabus’s AI policy. We anticipate that most library databases will begin using LLMs in some part of the information discovery or user guidance processes within the next year if they are not already. Any instructor of a class that requires students to conduct research should consider the role of emerging AI technology in library and publicly available information environments when developing an AI statement for their syllabus.

Below are a couple of sample statements related to library databases that you may consider incorporating into your syllabus alongside your existing AI statement.

If you permit students to use semantic search tools that use LLMs but do not want them to use generated content for their research:

The use of Large Language Models on the back-end of many library databases and online search engines has become commonplace. You do not need to cite instances in which you use an AI tool, such as a library database’s generative AI search feature or recommendation engine, to locate academic or other human-generated sources. The sources you find this way should be cited as you would any other source you may have found through traditional research methods. However, this exception does not apply to any AI-generated text, such as a generated research insight or research summary. You do not have permission to make use of AI-generated text in this class, even if that AI-generated text is found in a Champlain College Library database.

If you permit students to quote or paraphrase AI-generated text from a Champlain College Library database (particularly relevant if you differentiate between text generated within a library database with more quality controls vs. text generated by a more general tool such as ChatGPT; if you already allow the citation of any generated text, the statement below may be redundant).

You are permitted to quote or paraphrase AI-generated text from a subscription library database as long as you clearly indicate that the source is AI-generated and cite it appropriately. Library databases and other academic tools have safeguards in place to improve the accuracy of AI-generated text that are not necessarily present in other Large Language Model-based tools, such as ChatGPT. The generative AI tools in library databases can still make mistakes or have hallucinations in some circumstances, and you should verify any claims generated text yourself using human-created resources in the databases.

A note on why we are not offering suggested syllabi language banning the use of generative AI in all research contexts.

While the library and tools like Turnitin can always assist you in verifying whether material that a student quotes originated in a source created by humans, as well as if the sources students are citing actually exist and are not fabricated, neither the library nor any tool has the means of verifying whether or not a student used gen AI for information retrieval, provided they are properly citing human-created sources. In addition, some library databases are also using generative AI behind the scenes in ways that may interfere with the spirit of your policy, and it is often not transparent to us, let alone clearly labeled for students.

Further Reading

Communicating your ChatGPT/AI Policies from the Center for Learning and Teaching at Champlain College