AI Tools from EBSCO
Natural Language
The "Natural language" feature in EBSCO databases is an optional AI search assistant.
"Natural Language" is available in: Fusion and nearly 80 more EBSCO databases
The "Natural Language" tool interprets a conversational or natural language query, and translates it behind the scenes into effective keywords with Boolean operators (AND, OR, NOT).
This can be useful for complex queries, and when determining appropriate keywords is difficult.
Advanced Search Screen:
- Enable the natural language option on the Advanced Search screen by selecting the "Search options" tab below the search boxes, and then selecting "Natural language."
Basic Search Screen:
Select the "Natural language" toggle located beneath the search box on the right side.
Results Screen:
Select the "Natural language" toggle located beneath the search box on the left side.
Click the "Show refined query" link to view the keyword search that was constructed from your natural language entry.
In this example the Natural language AI tool constructed a search that includes additional keywords such as "lightning bugs" and "bioluminescence."
More information about Natural Language Search.
Artificial Intelligence (AI) Transparency Fact Sheet — Natural Language Search
AI Insights
AI Insights is is a document summary tool that highlights 2 to 5 key points for each document.
This can be useful to quickly review the relevance of search results.
"AI Insights" is available in: Fusion and nearly 40 more EBSCO databases
Tip: AI Insights does not display if the full text is not available within the database, or if the full text is too short.
- Click the "Generate AI Insights" button located at the bottom of each search result (when available).
After the key points are generated by the AI Insights tool, they will be displayed below the the AI Insights button.
More information on AI Insights.
Under the Hood
What powers EBSCO's AI tools?
EBSCO’s AI tools use Large Language Models (LLMs)—most notably Anthropic’s Claude 3 Haiku, made available via Amazon’s Bedrock API—for backend processing. These models are deployed for both the “Natural Language Search” (which interprets user queries and translates them into Boolean logic) and the “AI Insights” generative summary features. EBSCO also applies Retrieval Augmented Generation (RAG) methods to reduce hallucinations, and results are regularly evaluated using human-in-the-loop review for quality and accuracy.
AI-assisted, human-verified text
EBSCO Databases