While AI may seem to have grabbed the headlines out of nowhere, the latest buzz about tools like ChatGPT represents only one of the most recent - and perhaps most public – developments in AI technology. AI has been making impacts in many fields for several years now, and you have probably been benefiting from AI without necessarily being aware of it. A common example of AI that we have all likely benefitted from (or been frustrated by!) is predictive text.
As publicly available tools like ChatGPT bring AI to our attention, librarians have also discussed some observations about how AI is impacting library research and academic reading. There are already many examples of AI being used to improve existing search and discovery technology; for example, well-known resources such as the EBSCO databases are partnering with AI technology providers to improve search results (Expert.ai, 2021).
Another area where AI technology is being implemented is literature review software. There are several tools that use some aspects of AI LLM (large language models) and natural language processing to help map literature and point researchers to related content. While these tools are interesting to use for generating ideas and exploration of a topic, tools such as ResearchRabbit cannot be used for many forms of academic research and publication since the methodology of the tools is not transparent, and these tools rely on a limited (and often unclear) set of sources. As you have likely heard in your research methods classes, research has to be transparently reported, and good methodologies should be replicable. AI tools as they currently exist cannot be used for academic research because the way they make decisions and search for literature is not known. This is known as “the black box problem”; where the AI software itself and the developers working on it are not able to confirm exactly how the outputs and decisions are being generated (Blouin, 2023).
There are also now tools like Scholarcy that aim to use AI in order to help speed-up academic reading by generating summaries. Again, while an interesting concept, it’s not quite up to the level of critical reading that would be required in anything other than a casual reader. Elsevier (a prominent and prestigious academic publisher) has also made use of AI to create topic summaries, but crucially, Elsevier also ensures the AI-generated summaries are still edited and checked by subject experts.
As this handful of examples suggests, libraries and librarians are seeing that there is a lot to be gained by using the language-processing power of AI to help make academic research more streamlined, but the technology is not yet sophisticated enough to replace human assessment and critical reading, nor transparent and replicable enough to be used as a reliable academic research tool. It will be interesting to use this blog post as a “time capsule” to look back on as AI develops and academic publishing and research adapt to utilize and hone this technology!
Linked articles:
Blouin, L. (2023, March 6). AI’s mysterious ‘black box’ problem, explained | University of Michigan-Dearborn News. https://umdearborn.edu/news/ais-mysterious-black-box-problem-explained
Expert.ai. (2021, May 18). Expert.ai’s Natural Language Search Augments EBSCO Information Services’ Existing Search and Discovery Technology. https://www.expert.ai/expert-ais-natural-language-search-augments-ebsco-information-services-existing-search-and-discovery-technology/
0 Comments.