Generative AI : For Instructors

Generative AI and Instruction

Librarians will work with you to provide instruction on library resources, information literacy, and other information-related issues, which now includes generative AI. This instruction can look many different ways in practice, and we encourage you to work with subject librarians in your discipline to discuss. This page details shared principles that guide librarians' thinking and instruction around generative AI, as well as providing links to campus and other resources to help guide your own instruction on this topic. 

Generative AI and Librarians

The creators of this guide recognize that disciplinary lenses inform all of our perspectives on generative AI, and that librarians and other information professionals may notice and prioritize different aspects of this technology than experts in other domains do. Below are some ideas that guide librarians' thinking on this topic:

• Librarians are often associated with books, but we care about information, how it is disseminated, and how people become informed. 
• Librarians think of information-seeking skills as connected to, but distinct from, the skills required to create outputs, such as writing or presentations. The challenges that AI technologies present in regard to finding information are different than those one may encounter when approaching them as writing tools. 
• Unscoped generative AI tools appear to contain or produce information and appear to function like databases and search engines, but they have a completely different relationship to the body of text they draw upon to produce outputs. 
• Becoming an informed person is a skill that takes time and assistance from experts.
• We encourage solving information problems using the most appropriate, efficient tools rather than teaching to the tool. This may mean that what may first appear to be an information need best solved by generative AI is, in fact, a problem better solved by other means. 

Generative AI and Information Literacy

When meeting with students for information literacy sessions, librarians use the Association of College & Research Libraries’ Information Literacy Framework for Higher Education as a guide. This framework identifies threshold concepts, knowledge practices, and dispositions that guide library information literacy efforts across academic institutions. It includes six frames: 

• Authority is constructed and contextual
• Information creation as a process
• Information has value
• Research as inquiry
• Scholarship as conversation
• Searching as strategic exploration

How Generative AI Complicates the Information Literacy Framework

The existence of generative AI and, in particular, the promotion of generative AI technology as a means of seeking information introduce new complications into each of these frames. The following is a non-comprehensive list of areas in which generative AI and the frames intersect:

• As authorship requires intent, chat outputs are authorless. A key skill that librarians aim to teach students is using the context of authorship, publication source, and funding source to assess for authority. In the absence of this information, students cannot evaluate the output directly for relevant expertise and bias. 
• In understanding information creation as a process, students should learn to recognize characteristics of information products that indicate how they were created, including editorial processes and peer review requirements. Generated text has no such artifacts, and students are less able to make contextually appropriate decisions. 
• Students should understand information value in multiple ways, including commodification, intellectual property, and the ways in which systems that produce information marginalize some individuals and groups. With its impact on the journalism industry, copyright holders, and biased training data, generative AI touches all of these areas. 
• While we encourage students to develop their abilities to engage in research as inquiry and exercise strategic searching techniques, many generative AI companies market their products as simplifying search. Simplifying search misunderstands the exploration phase of the research process as a collection activity rather than a construction activity in which students develop their interpretive and creative abilities around the research topic.