Many library databases, including most of the ones below, automatically use now semantic (or natural language) searching. This means you can ask them questions like you would ask a person (or Google) and the results they return will not be limited to the keywords you use. The database's AI automatically understands the relationship between keywords and their synonyms.
Some library databases offer even more than that. For example, some offer "research assistants," which perform functions like summarizing articles, finding similar articles or recommending related terms. Currently, the following library databases have built-in AI research assistants. Watch the videos below for a description of what services their AI offers and how to use them:
Literature mapping is a way to identify academic articles by exploring connections between them. Articles can be "connected" by citations, authors, funders, keywords, and other means. These connections can be identified using free browser-based tools like the ones listed below:
How do literature mapping tools work? Watch this video demonstration of Connected Papers to find out:
Perhaps the most feature-rich literature mapping tool is Research Rabbit. Research Rabbit allows you to:
Watch the video below to learn how to use Research Rabbit:
Research assistance tools may perform a variety of services such as:
An AI research assistant that offers a free version is Elicit. Elicit gathers sources in response to a prompt in the form of a research question and can synthesizes these sources into a report. The user can then ask follow-up questions about the contents of the sources.
Watch the video below to see an example of how to use Elicit to find and summarize relevant research:
Another tool that includes AI research assistance is Semantic Scholar. Semantic Scholar provides free, AI-driven search and discovery tools, and open resources. It indexes over 200 million academic papers sourced from publisher partnerships, data providers, and web crawls.
Semantic Scholar finds relevant articles and suggests related articles. For select articles, it allows the research to query the article. Watch the video below to learn more:
As you develop your skills using generative AI, think critically! Here is a worksheet with some activities you might try:
Here are some additional ideas to help you think critically:
Try: In a library database, like Academic Search Complete, where "natural language search" is turned on by default, enter a question in the search box in the same way you would in Google (use a full sentence, don't worry about using discipline specific language etc). Try something like, "How does pollution in the ocean impact human health?" Once you get your results, click on the "Show refined query" link under the search bar and above your results. Examine the "subject terms" the database matched to your question.
Are they applicable/appropriate? Did the database suggest any matched subject terms you hadn't thought of?
Reflect: What are the advantages of library databases allowing natural language searches? What are some potential disadvantages?
Try: Use a tool like Research Rabbit to find sources.
Reflect: Where is Research Rabbit (and other, similar tools) searching for "connected papers?" Google that question if you don't know the answer. If you only used Research Rabbit, what sources would you never see? If you never used Research Rabbit, what sources might you miss out on?
What are some advantages of using Research Rabbit? What are some disadvantages?
What additional services does Research Rabbit provide that may be helpful?
Try: Find a paper in a traditional search, read it and summarize it yourself (write an overall summary, or focus on summarizing one aspect of it, like its conclusions or methodology). Then use a research assistant to find the same paper. Ask the AI to summarize the paper (again, focusing on an overall summary or one focused on a specific aspect).
Compare your summary to the AI's summary.
Reflect: How did the AI do? Did it miss any points that you included or include any points that you thought weren't relevant (it probably did)? Why might that happen? When is an appropriate use of this feature? What are the risks of this feature?
