How to search for information: AI in information searching

ChatGPT, Microsoft's Copilot, and Google Gemini are examples of conversational AI tools, which can answer questions, give translations and generate texts, instructions and programming codes for different needs. They are most useful when you already know something about the topic of your search and you are able to assess the text AI has generated.

Understand their limitations

• They can be helpful, but alone they are not sufficient for student's information searches.
• Many AI tools can make use of Internet searches when generating their answers. However, no generative AI is an actual search engine. The information they produce and the sources they provide should be approached critically.
• Check from other sources if the information generated by AI is true. AI also generates incomplete and false information and sometimes complete nonsense.
• It's not worth asking AI for in-text references or a list of sources, because
  • it often generates sources that do not exist
  • it can come up with in-text references with page numbers, but the information summarized in the text can't be found on those pages
• Even if an AI tool gives you a source that actually exists (website, article etc.), it might not be a good source to use in your academic studies. Remember to evaluate sources!

How to use generative AI in information searching?

Ask AI to suggest search words related to your topic. Don't settle for the first answer. For better results you can, for example

• request a list of keywords that cover different aspects of the topic
• ask AI to generate clarifying questions about your topic before suggesting keywords
• ask AI to tailor the search words to the information source, such as terms for searching for scientific articles in databases or finding information on the internet, etc.
• ask for keyword lists to be supplemented with synonyms and broader or more specific related terms

You can ask AI to generate search queries from the keywords or suggest databases where searches could be conducted. Remember to critically evaluate AI's suggestions, come up with additional  search words yourself, and refine search queries with the help of this guide.

Copyright and data security 

• Read the AI tool's  terms of use and the privacy policy. Find out how the service will handle the personal data you give when you sign up.
• Don't feed anyone's text or other material to AI tools without permission.
• Creating a summary or translation from a text of an e-book or article using AI tools may be convenient, but it may not be permissible. Before entering materials from library’s licensed databases into an AI tool, check the terms of use for licensed content. Enter licensed materials into AI if the terms explicitly permit it. You can find the terms of use for library-provided materials on the Databases A-Z list.
• Don't enter confidential information into any AI service, such as trade secrets, personal data, or similar sensitive details.

There are AI tools that are developed for students and reseachers and spesifically intended for searching research publications. Tools can be based on traditional keyword search (for example Semantic Scholar) or you can start your search by providing for example an article, its title, abstract or link. The tools then search for more articles on the topic and show connections between the articles. Some tools make visual maps of search results (for example ResearchRabbit, Connected Papers and Open Knowledge Maps), help identifying and defining key subjects in the text or create short summaries of the content and suggest research questions related to the topic (for example Elicit).

Keenious Plus available in Haaga-Helia

Keenious Plus is an AI tool which analyses text and PDFs to recommend relevant scholarly articles. You can use the tool through a web interface or an add-on for MS Word and Google Docs, which allows you to find literature suggestions as you write. The paid Keenious Plus has more features than the free Keenious version. To use the paid Keenious Plus, create an account with Haaga-Helia's email.

• Keenious guide

If you try these tools, bear in mind:

• The same copyright and data security considerations apply to these tools as they do to generative AI tools.
• Services cover varying amount of materials and fields of research.
• They usually work best in English.
• Not all sources found are freely available. You may be able to access the full text using HH Finna's article search. Search results also contain materials that are subject to charge.
• Some tools require registration. 
• Some tools are subject to charge or the fee-based version offers more features than the free version.
• The outputs of these tools should also be critically evaluated and verified. For instance, summaries generated from articles may not accurately represent the content.
• Traditional databases also have introduced AI-based features, such as additional material recommendations, AI chats or text translation features. New AI features are being be introduced, so keep an eye on those.

To find more tools and comparisons, you can search online, for example by using search words "AI research tools". You can search Youtube for demonstrations on different tools.