Skip to Main Content

Information Searching Guide: Information searching and AI

Use this guide to study the principles of scientific information searching: narrowing down a topic, choosing the right search terms and information sources and evaluating the search results. Mastering these valuable skills will prove useful in your studie

Using artificial intelligence to search for information

This page is about artificial intelligence (AI) in higher education. The content focuses on the use of AI in information searching, but we also address here general ethical issues raised by AI. AI and its applications are evolving rapidly, so this guide is a work in progress and will be updated when needed. So, visit the page for updates and follow developments in the field around the topic. 

AI is a broad concept that refers to the ability of computer systems to perform tasks and processes that normally require human intelligence. Such tasks may include image recognition, speech recognition, problem solving and decision making. AI is already used in many areas of our daily lives: within recommendation lists provided by applications, language translation, targeting advertising, predicting house prices, etc.

A company called OpenAI released its text-generating service ChatGPT for free in November 2022, and it became widely publicised in early 2023.


Key concepts

  • Artificial intelligence - Artificial intelligence is the ability of a machine to use skills traditionally associated with human intelligence, such as reasoning, learning, designing or creating.
  • Machine learning - Machine learning is a branch of artificial intelligence with its roots in statistics. Machine learning methods gives computers the ability to learn without explicitly being programmed. They improve their performance on a given task as more experience or data is accumulated.
  • Large language models - A large language model (LLM) is a model based on the probability of occurrence of words and word sequences. They predict the continuation of a given text input or produce the requested text.
  • Generative AI - Generative AI combines the power of machine learning, deep learning and artificial intelligence. It can generate original content such as text, video, audio, code or images in response to a given request. The generative AI model is trained using data and feedback and is then able to create ever more innovative outputs.

Tampere higher education community's guidelines

There are many guidelines and recommendations available on the use of AI, and ChatGPT in particular, in education. Some of the Tampere university community's guidelines are linked below.

Tampere university's student

Summary of the recommendations:

  • As a rule, the use of language models is allowed.

  • If a student uses a language model in an assignment or a thesis, for example, as part of language editing, this must always be mentioned.

  • Students are always responsible for the content of the texts they submit for assessment.

You can find the most recent Tampere University guidelines and recommendations in the Tampere higher education community's websites!

TAMK's student

Summary of the recommendations:

  • Students are advised to familiarise themselves with AI-based applications (such as ChatGPT) and use them as one of the tools in their learning.
  • Students are always responsible for the content of their own study assignments and the materials subject to evaluation.
  • TAMK requires students to use AI-based applications responsibly and ethically in their studies.

The most recent TAMK's guidelines and recommendations are in the Tampere higher education community's websites!

ChatGPT

ChatGPT is an OpenAI service that allows the use of different levels of language models, including ChatGPT (free) and GPT4 (paid), according to the user's choice. The application can be found at https://chat.openai.com/chat. To use it, you need an OpenAI account.

The service is used by entering text (prompt) and responding to a text prompt provided by the user. The service can therefore be used as a kind of dialogue on a topic of the user's choice.

ChatGPT has been trained on huge amounts of text and generates responses based on a statistical model. In simple terms, it can be compared to the predictive text input of a mobile phone. It knows nothing, has no access to the internet and cannot be used as a source of information. It is also unable to judge the reliability of the text it produces and can therefore claim things that are untrue. OpenAI has not disclosed any information about the data used to train ChatGPT. It is also worth remembering that AI models are notoriously biased.

Note! Open AI updated its paid ChatGPT-4 in September 2023, it is now able to use the internet.

Microsoft Copilot: an AI chatbot

Microsoft Copilot (formerly Bing Chat Enterprise) is a conversational AI application (GPT-4 language model and DALL-E 3 image model). Microsoft Copilot can also search for information directly from the Internet and can create images based on verbal descriptions.

Copilot is available for staff and adult students with a TUNI account. For students, it became available 03/2024.

The data processed by Copilot is not stored by it and the data is not used to train the AI. Moreover, the messages in the conversation with the AI are automatically discarded after the chat session ends. For more information, see Copilot's privacy policy.

See instructions on Copilot (intra). The Tip Bank article: Microsoft Copilot: Ten quick tips for use can also be useful.

ChatGPT for information retrieval

Orientation to the topic

  • You can use ChatGPT as a source of inspiration when brainstorming and formulating your research question. Be aware that it may introduce biased assumptions into the discussion. Remember that it has been trained with data up to 2021, as it has no knowledge of the latest research findings.

Planning your search

  • ChatGPT can help you find good keywords for your information retrieval. It is worth asking ChatGPT to complete its list of search terms several times, rather than settling for the first answer. Please note, however, that this does not replace the need to consult glossaries and check the terms used by experts in the field.
  • ChatGPT can help you identify tools or databases that are relevant to your topic. Please note that it may not provide a comprehensive answer from the available sector-specific databases. Nor can it tell you whether our library subscribes to a particular database. For comprehensive information on sectoral and cross-sectoral data sources, please consult the Library's Databases A-Z list.
  • Please note that ChatGPT is not trained to formulate search phrases for the purpose of searching for scientific information. If you ask it to construct search phrases for you, you might probably have to edit and reformat them.

Conducting a search

  • ChatGPT cannot help you with your actual search for information. It cannot search for information in a systematic and transparent way.   
  • ChatGPT can tell you how the database works. However, we recommend that you check the manual directly with the database of your choice. The functionality may have changed since 2021 and the ChatGPT response may well be out of date.

Evaluation of search results

  • You cannot use ChatGPT to evaluate search results. Information evaluation is a human skill based on criteria that you yourself consider relevant. As a researcher or student, as a data seeker, you have to decide whether a source is relevant to your work.
  • However, you can use ChatGPT to clarify and translate difficult texts, as it is good for summarising or reformatting text. However, remember to be critical of sources, i.e. don't believe everything - ChatGPT may present biased views.

References and citations

  • ChatGPT is unable to provide real sources for the texts it writes. It can provide a list of sources on request, but usually the sources are incorrect or completely fictitious.
  • ChatGPT is also unable to create bibliographies using certains citation styles. Reference management software, like Zotero and others, support a myriad of citation styles and are good tools for this purpose.

When using AI applications, always consider how they handle the data you enter. As a general rule, all data input to AI services is stored outside the EU/EEA and cannot be deleted from there. For example, no personal data, business secrets, confidential information or sensitive information may be entered into the services.

Instructions on how to restrict the use of your input for AI training in ChatGPT.

References to artificial intelligence

Should I refer to ChatGPT?

Yes, if you have used ChatGPT in your work, you must mention all its use. Check out the instructions of your university:

How do you make a citation to ChatGPT according to a specific citation style? See instructions e.g. : https://www.scribbr.com/ai-tools/chatgpt-citations/.

  • For example, an APA-style citation: 
    OpenAI. (Year). ChatGPT (Month Day version) [Large language model]. https://chat.openai.com

Writing your own texts (answers to assignments, essays, etc.) on ChatGPT is considered plagiarism or dishonesty in most institutions, even if you cite the source. Using ChatGPT in this way is not recommended.

Scopus AI

Scopus is Elsevier's abstract and citation database with a new generative AI add-on called Scopus AI. Scopus AI is based on the use of natural language, meaning that you can write your questions or hypotheses in normal language, regardless of keywords or Boolean operators.
Based on the question or hypothesis you entered, Scopus AI searches for publications published since 2013 and uses their abstracts to create an answer, i.e., a short summary of the topic. The summary includes references to articles found in the Scopus database. In case you want more in-depth knowledge, you can continue exploring with the help of an extended summary, concept map, basic publications in the research area, and expert sections on the subject. 

Artificial intelligence in academic information searhcing

Although applications of generative AI (such as ChatGPT) are not, at least for the time being, suitable for scientific information retrieval, there are a number of different applications that use AI specifically for article retrieval. You can find some of them listed below.

NOTE! The services presented below are not organization-licensed and trial and use of the services is at your own risk. Remember information security (see. Artificial intelligence (AI) and information security). Please also note that services may become chargeable or disappear.

AI-powered article search tools

  • A free and a paid version are available. The paid version, Keenious Plus, includes more features than the free version.  Once logged in, you will have access to more features.
  • A tool that uses an existing text as a basis for information retrieval. It searches for articles on the same topic based on an existing article or other text. In addition to the article search, Keenious generates a subject list, which it interprets as related to the text that was fed to it. Based on this, the searcher can refine his/her search further.
  • This tool can be used not only on a webpage, but also via a Word or Google docs plug-in.
  • So far freely available, but it is unclear how long this possibility will remain. Once logged in, you will have access to more features.
  • The search is similar to a word search on Google Scholar.
  • Displays article references and tries to find, not only the most referenced articles, but also the most relevant ones. It also offers article recommendations based on the articles selected by the user.
  • Suitable for users who are interested in filtering and sorting search results and saving articles for new article recommendations.
  • There is no charge to start using the service, but registration is required. The service becomes chargeable once a certain amount of usage is exceeded. The amount of use determines the price.
  • This tool searches and summarises scientific articles based on a given research question or an article. On the search results page, Elicit does not display the number of results (like traditional search services) but shows the top seven most relevant articles.     
  • Elicit also provides other tools for academic work helping you to formulate your research question, analyse articles and draft abstracts.

Visualising academic information retrieval tools:

  • Free and paid versions are available with differences in the number of articles and maps.
  • Creates a bibliographic map with data from a single publication or, alternatively, from multiple publications using BibTeX/RIS files.     
  • Litmaps shows the relationships between scientific articles in a dynamic network graph format. It makes suggestions for articles based on the relationships between articles, which helps you to find articles outside keywords.    
  • Requires you to create a login.
  • You can start using the tool by adding one good article to the collection. The tool generates recommendations based on your collection using titles, abstracts and other metadata of the articles.   
  • Research Rabbit visualises the networks formed between publications. You can move forward on the map, creating new publication networks including authors and sources. The tool helps to find new connections between a certain phenomena being studied and the researchers.

Language models combined with search engine

Linking the language model to the internet and a search engine allows the language model to provide answers based on internet searches. For example, ChatGPT+plugins, Microsoft Copilot (former Bing Chat), Google Gemini (former Bard), Perplexity.ai and You.com.

Operating principle:

  • The language model interprets the user's question and possibly rephrases it.
  • According to the understanding of the model, suitable search words are entered into the search engine through the software interface. The language model interprets the user's query and possibly reformulates it.
  • The search engine returns the best results for the language model.
  • The language model answers the question based on the sources it receives.

Even in this case, the language model may be hallucinating or misinterpreting sources. So, remember to be critical of the sources. Most tools require registration.

Ethical and evaluative use

Generative AI is a tool that can help us in our daily lives, at work or in our studies. As with any other tool, ethical, evaluative and appropriate use is the key point.

  • Integrity
    • Can AI be used in a mission? If so, what for? And how should it be reported?
  • Mis- and disinformation:
    • Misinformation is incomplete or incorrect information, is given inadvertently and is not intended to mislead.

    • Disinformation is the deliberate distribution of false information, which can be motivated by, among other things, political and social influence, financial gain and malicious intent.

  • Information literacy
    • AI tools do not necessarily report the original sources they use, nor do they necessarily use sources that meet the requirements of scientific writing. In some cases, these sources have proven to be non-existent or inaccurate. If you use a tool that produces some elements of your work, the person reviewing the work needs to know what parts are yours and what is from elsewhere.
    • AI-powered search tools may seem to make scientific research easy and fast. However, from the student's perspective, this can mean that the student is not learning important skills, such as the basic skills of scientific research like information retrieval, critical evaluation and problem solving.
  • Protection of privacy
    • For example, to use ChatGPT, you have to create an account, which allows for data collection. This is a privacy issue. AI tools may require you to provide a phone number or other personal information. Users should be careful about what information they share when creating an account on the web.
    • Please read the program's privacy policy and terms of use carefully. Remember that when you use the tools, your conversations are typically stored by the tool and used for that tool's purposes.
  • Authors' copyrights
    • In some cases, AI tools can use online material without the permission of the authors. Respecting copyright is a part of academic good scientific practice, which is why it is very important to be critical of the use of such tools in academic work.

Checklist for students

  • Discuss with your course teacher or seminar tutor well in advance whether (and which) AI applications are allowed for information searching for your work.
  • Familiarise yourself with AI applications beforehand: what they can and cannot do to support your search, compare the results.
  • Consider what you can and cannot use an AI application for, when searching for information.
  • Learn about the AI application you are using: who is running it and why; is it fee-based or restricted, does this affect the results?
  • Source criticism is still your responsibility. Be critical of the results: do comparative sampling of different sources of information and check, for example, the references provided by the AI and the citations they get from databases.
  • Cite AI as instructed if you use it.

Guides and guidelines from other libraries

Read more

Rudolph, J., Tan, S. & Tan, S. 2023. ChatGPT: Bullshit spewer or the end of traditional assessments in higher education? Journal of Applied Learning & Teaching. 6(1).https://doi.org/10.37074/jalt.2023.6.1.9 

Teel, Z., Wang, T., & Lund, B. 2023. ChatGPT conundrums: Probing plagiarism and parroting problems in higher education practices. College & Research Libraries News, 84(6), 205. https://doi.org/10.5860/crln.84.6.205 

What is AI?

What is AI? A video by Museum of Science, Boston (4,38)

Logo

Email: library@tuni.fi
P. 0294 520 900

Kirjaston kotisivut | Library homepage
Andor

Palaute | Feedback