Dean of SBU Libraries Discusses Future of AI
Karim Boughida, dean of Stony Brook University Libraries, has been selected as the incoming chair of the National Information Standards Organization (NISO), transitioning to the role after serving as vice chair of the organization. NISO develops, maintains and publishes technical standards related to publishing, information, and library applications.
Boughida joined Stony Brook in September 2022, and is working with his team to reimagine the vision, mission, and strategic plan for the library to be bold, innovative, and leaders in advancing research and learning; to think beyond of the library as a means of support and to introduce the concept of the library as the interdisciplinary hub of the university.
Part of the plan to be innovative is to create a space to discuss artificial intelligence (AI). In his former position as dean of University Libraries at the University of Rhode Island, Boughida spearheaded efforts to create the first AI Lab in a library worldwide that focused on AI literacy and the ethics, social impact, and social justice issues surrounding AI. The lab also provided guidance and recommendations on issues like how to use AI in daily life.
Boughida hopes to create a similar hub at Stony Brook to discuss these important issues and to integrate AI into the library as a means of assisting users and democratizing key concepts like machine learning, natural language processing, and robotics. Stony Brook has a strong record of AI research and the focus on AI within the university libraries aligns with the priorities and focus of the university.
“Typically, when you say that we have an AI Lab, people think that it is housed within an academic department, and even if you say that it is open to everyone, people outside of the department do not feel as welcome. When it’s in the library, because we serve everybody, it is democratizing AI,” Boughida said.
Boughida saw students and faculty as well as community members utilizing the services of the AI Lab. Small businesses who did not have the means to hire an AI specialist visited the lab to find out how they could use AI to improve their business, and the lab became a resource for economic development within the community.
“We are also discussing how we can improve services within the libraries. Can we use a conversational agent as the first step for library services? Or a smarter chatbot? How can we use AI to increase and have a better search in terms of our library catalog?” said Boughida, who noted that the libraries are seeing an increase in users searching for a scientific article that does not exist, but was cited in an AI-generated piece. “The big problem we face now with generative AI is that AI is not in the business of truth or facts. It provides you with a likely combination of word following another word following another word, and provides users with 40 to 70 percent of citations that are fake and do not exist. We have to educate these users on what is real and what is fake.”
“You can get good content from AI, but you have to parse it, and figure out what is real and what is fake,” he added. “And that is where AI literacy is so important.”
Boughida described how faculty are adjusting to the ‘new normal’ in which students are using AI generated text, by assigning a greater number of classroom-based assignments, without the use of technology, or asking students to combine results from generative AI with their own thoughts.
Boughida plans to soon hire a director of AI and data science for the libraries, along with a data scientist and a natural language processing specialist, to ensure that Stony Brook libraries will remain leaders in advancing research and learning.
“We need to adjust our pedagogy and data, since detectors are not reliable at catching AI-generated texts — this is a new normal. It’s happening, so we need to work around it and use it as an enhancer for learning,” he explained. “It’s going to change our lives, so that’s why we have to be a part of it. AI can be used for both good and bad, and is inherently not bad or good. Hollywood and some longtermists manufactured the image of AI by showing AI as villains, and we need to combat these messages and show the potential for good.”
— Beth Squire