The article reports on an exploratory study that assesses the results produced by emerging artificial intelligence (AI)- and large language model-driven search tools in response to a series of queries and prompts based on four scenarios of information-intensive tasks of university students and researchers. Sixteen questions and prompts were created based on four scenarios of information-intensive tasks of university students. Each of these questions and prompts was presented to six AI-driven search tools, and the results were manually checked to assess their suitability for specific user needs and contexts. Based on the findings, it was argued that while the AI-driven tools bring a paradigm shift in information access for education and research, outputs generated by these tools vary quite significantly. Choice of the right tool, framing the question and further prompting play a key role. Also, users need to scrutinise each output to check their quality and reliability in the context of the specific search tasks. It was concluded that further research is needed involving different user groups, scenarios and search tasks and different AI-driven search tools. Implications of the use of AI-driven search tools for libraries and scholarly databases, as well as for research and scholarship in different areas of information science, are discussed.
Read full abstract