ChatGPT, Ethics, and DEIA Intersect at ALA Annual 2023

Considering the implications of AI ethics for DEIA

Posted on June 28, 2023 in Blog Posts

The 2023 ALA Annual Conference & Exhibition recently concluded in the Windy City, where it was host to an impressive array of exhibits, discussion forums, author signings, and sessions, covering everything from book bannings to adult literacy and, what is likely of most interest to LTI readers, emerging trends in library technology. Although not featured as the central topic of most technology-related discussions—at least based on a cursory survey of event offerings (see chart below)—a hotly discussed subject within the library technology sphere was AI and its implications for librarians. 

Many articles published in the last year have highlighted the ethical concerns surrounding generative AI like ChatGPT. These same concerns were echoed across conference events but were considered alongside the potential opportunities that this type of technology presents, enriched by poignant reflections on the implications for diversity, equity, inclusion, and accessibility (DEIA). 

---

🌟 Want to read more from Toward Inclusive Excellence? Check out:

• The Life and Times of Diamond and Silk Underscores the Tragic Human Impact of Disinformation
• We Must Dismantle Misogynoir and Allowing Only One BIPOC Woman to “Win” at a Time
• The Faux Pop-Feminism #GirlBoss “Movement” Imploded Largely Due to Racism and Classism

---

Two sessions in particular gave rise to fruitful dialogues incorporating DEIA. The first, a discussion forum on access, accessibility, and emerging technologies hosted by the Emerging Technologies Section of the Reference & User Services Association (RUSA), brought together public and academic librarians to contemplate how chatbots and other forms of AI can improve overall accessibility in libraries. “AI-powered tools are one way to improve accessibility,” as Hong Zhou and Sylvia Izzo Hunter detail in a guest post for The Scholarly Kitchen, referenced during the forum. Examples discussed at the event include Seeing AI, which helps visually impaired users access written or visual content, and Speechify, a text-to-speech extension that can even be paired with ChatGPT, the widely debated large language model.

As librarians at the conference contended, ChatGPT has its benefits for those who struggle to express themselves through writing and can even be an effective tool when used as a springboard to generate ideas. However, for reference purposes, ChatGPT is still very limited. A common concern raised several times at the conference: ChatGPT cannot yet cite sources and will often create fake citations if asked for references. 

This is a major concern that was also addressed during the Core Top Ten Technology Trends panel discussion on ChatGPT, where panelists stressed the importance of librarians’ learning the ins and outs of AI like ChatGPT, given that students will inevitably use the program. To drive home this point, Hannah Byrd Little, director of library and archives at The Webb School in Tennessee, cited an opinion piece for the Chronicle of Higher Education in which Columbia University undergraduate Owen Kichizo Terry details how students can use ChatGPT, not to write whole papers but to seek out sources, asking the AI for information and writing essays based on what it tells them, possibly without validating that information. 

This raises many anxieties around both academic integrity and critical thinking, but it also presents an opportunity for librarians to become experts on ChatGPT in order to teach patrons, especially students, how to use the chatbot smartly and effectively. This seemed to be the red thread running through both AI-centered discussions: librarians’ key role in fostering the ethical use of AI, which could otherwise have many negative downstream effects if used without guidance.  

Without appropriate instruction, for instance, users could become too trusting of the information generated by an AI like ChatGPT, treating it as an authority for generating statistics or demographic data, which could even be used to make policy decisions regarding certain communities—a chilling scenario outlined by Jonathan McMichael, an undergraduate success librarian at Arizona State University. In such a case, AI could become a tool for dehumanization. This concern was seconded by Trevor Watkins, a teaching and outreach librarian at George Mason University, who pointed out how AI is often weaponized, as in the case of surveillance. ChatGPT could be similarly weaponized, he contended, for instance to make authoritative statements that are harmful to minority communities or to generate hate speech.  

This fear is not unfounded. In a June 22 interview with Democracy Now!, Dr. Joy Buolamwini of MIT’s Media Lab and co-founder of the Algorithmic Justice League spoke on the documented racial and gender biases embedded within AI. This is because emerging AI systems work off of pattern recognition, drawing from large datasets to glean information. Because knowledge production is never neutral, the data themselves can be biased against certain groups. Given these concerns, there has even been talk of creating a new federal agency to monitor and regulate AI, though not much progress has been made on this.  

If librarians can take away anything from ALA’s 2023 annual conference regarding the growing influence of AI, it is that AI must be approached through collective action. As stewards of knowledge and advocates for information literacy, librarians are ideally poised to instruct their patrons on the appropriate uses of AI, to ensure that there is representation in the sources being consulted through AI, and to encourage users to always think critically about the information they consume. 

---

👯 Want to learn more about DEIA in higher ed? Check out our sister blog, Toward Inclusive Excellence.

🔥 Sign up for LibTech Insights (LTI) new post notifications and updates.

✍️ Interested in contributing to LTI? Send an email to Deb V. at Choice with your topic idea.

---

Related Posts