Using AI to support information retrieval
Article sections
Artificial intelligence offers many opportunities to enhance the information retrieval process in higher education. This guide provides examples of how AI can be used in different stages of information searching.
Language model-based AI chats, such as ChatGPT or Microsoft’s Copilot, are not suitable for actual professional or scientific information searching but are helpful in planning search and brainstorming keywords.
AI chats may provide you with links to sources, but the sources might not be reliable, and you won’t be able to evaluate why those particular sources were selected from all the information available online. Therefore, be critical when using AI chats.
For actual scientific information searching, it is advisable to use AI applications specifically designed for searching information from scientific databases.
1. Formulating the research question with AI Chats
- AI chats like ChatGPT or Copilot can assist in refining and structuring your topic. They help you consider your topic from different perspectives.
- Input your topic into the language model and ask it to formulate your information retrieval question. Request the chat to ask you clarifying questions before responding.
- Example prompt: “Formulate a research question on ‘the effects of climate change on agriculture in the Nordic countries.’ Ask me clarifying questions before responding.”
2. Brainstorming Keywords with AI Chats
- You can also use language model chats to brainstorm keywords for searching information with search engines.
- Input your topic or research question into the language model and ask it to suggest keywords. You can refine your question by specifying what kind of information you are looking for.
- It is also advisable to ask the chat to pose clarifying questions before responding. This helps you refine your information retrieval question from different perspectives.
- Example prompt: “Suggest keywords for searching research information in Finnish and English on ‘principles of sustainable development in corporate management.’ Provide also specific keywords. Ask me clarifying questions before responding.”
3. AI Applications Designed for Scientific Information Retrieval
- AI applications designed for scientific information retrieval search sources from scientific databases. Using them complements traditional information search.
- Benefits of using AI applications designed for scientific information retrieval include a multidisciplinary approach and visualization tools. Many applications also allow you to search for information with a naturally written question.
- These tools often provide various options for analyzing the found sources.
- NOTE: Copyright also applies to scientific articles. Therefore, do not input articles or other texts written by someone else into the AI application without permission.
- NOTE: Some tools require creating a separate user account. Before creating an account, review the service’s privacy policy, as most applications do not handle personal data according to European privacy regulations.
Examples of AI applications for scientific information retrieval
Consensus
- You can search for information by writing a research question in natural language.
- Produces a summary of the found sources with links to the sources.
- You can limit the search to peer-reviewed sources only.
- Requires account creation, limited free use.
- Paid version offers more extensive usage options.
Elicit
- Search works with a research question, concepts (keywords), or a PDF file.
- Summarizes articles and allows you to ask questions about the articles.
- Requires account creation, limited free use.
- Paid version offers more extensive usage options.
Keenious
- Search works with a PDF file, URL, or text (keywords, research question).
- Provides information about citations and references of the article and recommends similar articles.
- No mandatory login, additional functions with free login.
- Paid version offers more extensive usage options.
- Operates within the EU and is GDPR-compliant, does not store submitted texts.
Litmaps
- Search works with keywords, author name, DOI address, etc.
- Creates a map from the selected article showing citations and references and suggests similar articles.
- No mandatory login, additional functions with free login.
- Paid version offers more extensive usage options.
Research Rabbit
- Search works with keywords, author name, DOI address, etc.
- Creates a map from the selected article showing citations and references and suggests similar articles.
- No mandatory login, additional functions with free login.
- Paid version offers more extensive usage options.
Semantic Scholar
- Search works traditionally with keywords, similar to Google Scholar.
Provides information on citations and references of the article and recommends similar articles. - No mandatory login.
Author: Minna Kivinen