# Better and worse ways to use AI for information retrieval

> A guide on how to effectively and responsibly integrate AI tools into the information retrieval and research process.

## 1. AI as a Tool for planning and understanding information retrieval

*   Formulate a research question or topic
*   Explore perspectives and generate ideas
*   Clarify concepts (verify them from reliable sources)
*   Test your understanding and structure the topic
*   Use AI as support while making evaluations and decisions yourself

AI can support brainstorming and organizing ideas, especially at the beginning of the search process. Responsibility for evaluating and using information remains with you.

## 2. Generating search terms and concepts

*   Identify search terms, synonyms, and terminology
*   Broaden searches from different perspectives
*   Test and refine search strategies

AI can help generate search terms, but subject thesauri and disciplinary vocabularies are also important sources.

## 3. Using AI Assistants in scholarly databases

*   Use AI tools built into databases
*   Critically assess their summaries and suggestions
*   Explore citation networks and topic maps critically

Unlike general generative AI, these tools search only within database content, helping to locate reliable information.

## 4. AI-Based scholarly discovery tools

*   Find publications relevant to your research question
*   Explore topics through citation networks and visualizations
*   Particularly useful at the start of a search
*   Often partly free to use

These tools do not find everything. A comprehensive literature search must still be carried out in scholarly databases.

## 5. Critically evaluating results

*   Verify information from original sources
*   Assess source reliability
*   Consult multiple sources

AI-generated information must always be checked against the original source and cited accordingly.

## Worse ways to use AI (What to avoid)

### 1. Using AI as an information source
*   Copy AI-generated answers directly
*   Cite AI as a source

AI-generated content should not be used as a source because its original sources may be unclear and AI does not create new knowledge.

### 2. Accepting sources without verification
*   List articles suggested by AI without checking them
*   Fail to verify whether sources or articles actually exist

AI may suggest irrelevant or even non-existent sources that appear credible.

### 3. Using AI for the actual literature search
*   Let AI conduct the search for you
*   Try to find scholarly sources using only ChatGPT
*   Skip databases and discovery services

Generative AI tools are not designed for reliable information retrieval, and their results always contain gaps.

### 4. Trusting AI responses uncritically
*   Assume AI-generated information is correct
*   Do not verify content from original sources
*   Ignore errors and biases
*   Fail to evaluate sources and content yourself

AI can produce inaccurate, biased, or misleading information. It can support your work, but it cannot replace your own expertise.