Artificial intelligence is increasingly recognized as an essential new frontier in higher education. But both its capacities and its applications are evolving so quickly that it remains unfamiliar and uncertain terrain for many scholars. That’s why University Libraries is delighted to welcome a new librarian, Applied AI-Data Science Specialist Vandana Srivastava, to our Digital Research Services Team.
As a Ph.D. candidate in Computer Science and a member of USC’s AI institute who also has years of experience as an instructor and an analyst, Srivastava has both the knowledge and the perspective to help members of the USC research community navigate the possibilities and the pitfalls of AI. In fact, says Director of Digital Research Services Kate Boyd, “because she’s a Ph.D. candidate, Vandana’s been in the same situation the people she’s helping are in: of finding out the best ways to use AI in research. And she has technical knowledge that really enhances the support the libraries can provide to researchers.”
Having an AI librarian will, Boyd says, significantly enhance the Libraries’ ability to support researchers across all disciplines who are working with large data sets or who could benefit from using various types of AI. “She brings serious math, statistics and data skills that the Libraries can really benefit from,” Boyd notes. “And having her here is a boon for the USC research community, because we’re one of the few academic libraries that have an AI librarian.”
We asked Srivastava to explain how she envisions her role and how she can contribute to the Libraries’ support of the research enterprise at USC.
What kinds of expertise do you bring to this role?
As an experienced practitioner of AI, I can guide people not just on what AI tools can do but how to use them in conducting research- for example, how to create prediction models for textual data. Many researchers are working with very specific kinds of data sets, and I know a lot about which AI programs will work best with which kinds of data. I also know how to find the right information when people are working on things that are very specific or esoteric. And if I’m not the right person to help with a specific need or problem, I can direct people to someone who can.
You have a background in both industry and academia. What drew you to AI librarianship?
I have always loved to teach and to mentor. I’ve done a lot of mentoring of other women in tech and find it very rewarding to help people gain confidence with new technologies and their applications. And as a member of the USC community, I’ve had great experiences working with University Libraries, so I knew working with the Digital Research Services team would be a good fit.
How can members of the USC community use your services?
I’m available to help at all stages of the research process, across any discipline, and where AI methods are needed for analysis. My very first consult was with a graduate student in history who was a prime example of how I can be of assistance to people in fields in the humanities who have the vision to tackle a data-driven research project but not the training.
What do you find most exciting about AI as a research tool?
AI is actually very simple: it’s about input leading to output. So why is it creating such a buzz? Largely because of its capacity to find complex relationships between input and output. And it is always learning and increasingly able to understand more and more complexities.
What are some of the biggest advantages – and pitfalls – of using AI in an academic setting?
AI does two major things that can be very useful across multiple disciplines: prediction and classification. It’s also helpful with coding, which can take a long time for occasional coders or non-coders. But if you rely too much on it for your academic work, you do not retain the knowledge that comes by reading and practicing. In short, it’s a very good tool, but there are ways in which it cannot replace the human touch. We are all unique as humans, and that uniqueness is what makes our scholarship distinctive.
If a USC faculty member or student wants your help with their research, what’s the best first step?
Just email me: vandana@email.sc.edu. We’ll set up a time to meet so I can learn more about your project and your needs, and then we’ll take it from there.