Skip to Main Content

AI (Artificial Intelligence): AI in our Databases

Responsible and Effective Use of Artificial Intelligence in Academic Research and Writing

AI in Our Subscribed Databases: What You Need to Know

This page provides an overview of the current AI-powered features in the Unisa Library databases, grouped by functionality, with practical tips and critical notes on reliability and ethics.

Databases that have AI integration will have a AI assisted flag flag next to the database name on our A-Z list. 

Scope:

  • Covers AI features currently available in subscribed databases.
  • Focuses on research-related tasks (searching, summarizing, recommending).
  • Includes vendor-provided tools only (not third-party add-ons).

Important: AI features are vendor-driven and may change. Always verify outputs against original sources.

Extra reading: Taylor, J., Dagan, K., Youngberg, M., Kaufman, T., & Radding, J. (2025). A Survey of AI tools in Library Tech: Accelerating into and Unlocking Streamlined Enhanced Convenient Empowering Game-Changers. Journal of Electronic Resources Librarianship, 37(2), 217–229.

AI features

What It Is:
How to Use:
  • Use full sentences or questions for broader results.
  • Combine semantic search with Boolean operators for precision.
Be Aware: 
  • Prompting may introduce bias.
  • It is sometime difficult to ascertain what the ranking logic of the results are.
  • There is usually little transparency in how the AI interpreted the search query or prompt, making it difficult for users to adjust their searches for more relevant results. 
Example of Database Using it: Summon Discovery Service

Important: AI features are vendor-driven and may change. Always verify outputs against original sources. AI outputs should never replace critical reading or citation.

What It Is:

An AI chatbot that provides synthesised answers and citations 
How to Use: Getting a quick, cited overview of a complex topic. 
Critical Notes:
  • Chatbot responses may lack transparency and can include inaccuracies, so always verify information against original sources.
  • Natural Language Search uses an LLM to convert the user’s query into a Boolean search string.
  • The Boolean string is hidden from the user and then processed through the usual search algorithm and index.
  • Because users do not see the Boolean query, results may appear unrelated to their original natural language terms.
  • This lack of transparency makes it difficult for users to refine searches or correct mistakes.
  • Avoid copying AI-generated text into academic work without proper validation and citation.
Example of Database Using it:

Important: AI features are vendor-driven and may change. Always verify outputs against original sources. AI outputs should never replace critical reading or citation.AI outputs should never replace critical reading or citation.

What It Is:

Database suggests additional keywords after the first search, from the search results page. 
How to Use: Broadening or narrowing of a search strategy interactively. 
Be Aware:
  • Suggested keywords may reflect algorithmic bias, so verify their relevance before adding them.
  • Over-expanding your query can introduce noise and reduce precision; always balance breadth with specificity.
  • AI-generated keywords may not align with controlled vocabularies or indexing standards, which can affect systematic searches.
Example of Databases Using It:

Important: AI features are vendor-driven and may change. Always verify outputs against original sources. AI outputs should never replace critical reading or citation.

What It Is:

AI generates summaries or highlights key points from articles.
How to Use:
  • Use summaries for initial screening, not for citation.
  • Always read full text for nuanced arguments.
Be Aware:
  • Users often can not tell how much an AI’s answer is based on the original source.
  • AI tools handle sources in different ways, so outputs vary.
  • The output might be:
    • an exact copy of the text
    • a summary of key ideas
    • a list of possible research topics
    • something completely different
  • These summaries can be wrong, especially for creative or unusual writing.
  • Do not copy AI-generated text into academic work without verification.
Examples of Databases Using it: 

Important: AI features are vendor-driven and may change. Always verify outputs against original sources. AI outputs should never replace critical reading or citation.

What It Is:

Provides takeaways, related topics, suggested sources, and indexing terms. 
How to Use: Deeply analysing a key paper to fuel a new search strategy; literature mapping. 
Critical Notes:
  • AI research assistants do more than check sources: they can bypass much of the idea development and refinement process.
  • They generate large amounts of text that can not be reproduced, raising concerns about academic honesty, authorship, plagiarism, and attribution.
  • AI summaries and suggested sources may:
    • oversimplify complex arguments
    • introduce bias, so always verify against the full text
  • Suggested indexing terms might not match controlled vocabularies, reducing precision in systematic searches.
  • AI-based literature mapping can:
    • miss key foundational works
    • overemphasize recent trends. So use it only as a starting point, not a final strategy.
Example of Databases Using it: 

Important: AI features are vendor-driven and may change. Always verify outputs against original sources. AI outputs should never replace critical reading or citation.