Since its launch in late 2022, ChatGPT has inspired much discourse on potential gains and perceived pitfalls surrounding its use, particularly in academic productivity and scholarly research. The generative artificial intelligence (AI) product created by OpenAI offers responses to questions posed by users, and is trained via machine learning and other language models to provide more refined conversations over time. A new iteration called GPT-4, which can read imagery in addition to text inputs, promotes “safer and more useful responses.”  

Capabilities and limitations provided by ChatGPT

The implications of ChatGPT and other generative AI platforms on information literacy and academic research are part of ongoing discussions at University of Houston, in which faculty and librarians are familiarizing themselves with the technology and gaining a deeper understanding of how students are using it. In conversations with students and faculty, librarians in the department of Teaching and Learning at UH Libraries have learned that students are interacting with ChatGPT for assignments in a variety of ways. They shared that a common use is to prompt ChatGPT to write an essay or at least provide a starting point. Students are also asking it to generate summaries of articles and books to supplement their learning, gain foundational knowledge about a broad topic, or find potential sources for research papers. 

The academic and scholarly utility of ChatGPT is variable. Because it sometimes provides erroneous or false responses, relying on it to produce a rigorous essay or provide sources that meet academic standards is risky. The temptation for some students to take ChatGPT’s responses and submit as their own work can be overwhelming. It’s a widely available tool that is here to stay, however, and knowing how it operates, as well as how to use it appropriately, will help mitigate those drawbacks. This is where information literacy plays a significant role in supporting academic integrity, creating an environment in which generative AI and information literacy can coexist for the benefit of scholarly users. 

Information literacy involves “the reflective discovery of information, understanding how information is produced and valued, and ethical participation in communities of learning.” Teaching and Learning librarians view ChatGPT and other generative AI “as an opportunity to emphasize the importance of verifying information and reinforcing that as a major part of the research process.” Moreover, AI models offer the potential to assist students with language or communication disabilities in drafting or improving essays and other writing. Students might also use ChatGPT to improve vocabulary, sentence structure, and overall communication when learning a new language.   

UH librarians are responding to the use of ChatGPT by developing their knowledge of generative AI through testing, and are collaborating with UH faculty members to identify teaching materials that could provide instruction for students on applying ChatGPT ethically and as part of an iterative process. Instruction@UH, sponsored by the Office of the Provost, is an online resource for faculty to engage in areas of educational technology and instructional design. A category of content dedicated to AI and ChatGPT offers discussion and practical advice for instructors, such as recommendations on how to integrate the emergent technology in coursework. 

For this feature, ChatGPT was asked “how does ChatGPT relate to information literacy?” Its partial response: “ChatGPT, as a language model, can play a role in supporting information literacy…. However, it’s important to note that while ChatGPT can be a helpful tool, it’s not a substitute for developing one’s own information literacy skills. Users should critically evaluate information obtained from any source, including ChatGPT, and seek multiple perspectives to form well-rounded conclusions.”  

Read ChatGPT’s full response to the information literacy question.