EBSCO Information Services pursues Generative Artificial Intelligence opportunities

Company embarks on projects to determine how GenAI can enhance search discovery and content creation while avoiding "hallucinations" and spurious information.

EBSCO Information Services announced it is embracing the power of generative artificial intelligence (GenAI) because it recognizes the transformative potential of generative AI in the realm of academic research and libraries. It has several AI pilot projects in specific environments with the goal of amplifying the effectiveness of research. In the press release,  EBSCO reveals its lofty goal—to set a new standard in information discovery and comprehension for students, faculty and researchers.

EBSCO has been investing in a modern user interface (UI) for the company’s research platforms, including EBSCO Discovery Service (EDS) and EBSCOhost®. The company’s AI pilot projects will integrate into the modern UI and include natural language search, improved evaluation of search results through summaries and key insights, conversational interactivity with source documents — including multilingual — as well as search alerts and content recommendations. These insights will be clearly labeled as AI generated, and if desired, the library administrator may remove the feature from the library's profile.

Having used AI operationally for a number of years, EBSCO is expanding with GenAI to enhance metadata, categorize content and build linked data relationships. With linked data, generative AI models can access a vast network of interconnected information, including vocabularies and knowledge graphs. This approach will help with overall computational costs as well as the environmental impact of generative AI, while ensuring that overall adoption is not resource-heavy for libraries.    

While generative AI holds remarkable promise, it's crucial to note its vulnerability to potential misinformation and factual inaccuracies. With EBSCO's mission to serve researchers with reliable, credible information, it is adopting a meticulous, quality-centric approach. By embedding AI capabilities within the framework of its rigorously curated content, EBSCO hopes to not only help prevent the dissemination of misinformation but also ensure that its AI technology delivers highly relevant and accurate scholarly resources, augmenting the reliability and precision of research across the academic landscape. Pilot projects will focus on finding the most affordable, high-impact options for libraries.