MetSearch Research Assistant is an innovative search tool that leverages AI technology to search Cardiff Met's library catalogue and search system used by Cardiff Met.
Core utilities
The tool offers a number of key functionalities that can simplify advanced search techniques:
- Natural language querying: Students can ask research questions in everyday language rather than requiring Boolean search expertise. It achieves this by converting natural language questions into Boolean search strings.
- AI-generated summaries: Provides synthesised overviews with inline citations from the top 5 most relevant sources. Provides clickable links to sources and full search results
- Source recommendations: Suggests 5 specific academic resources from the Primo Central Discovery Index (CDI).
- Related research questions: Suggest directions for further research by offering related questions to expand research scope. Also allows user defined search query refinement.
- Quality filtering: Automatically excludes newspapers and retracted documents.
- Multilingual support: Available in over 50 languages.
- Integration: Embedded within the Cardiff Met Primo library catalogue system.
Benefits and advantages
This Open University report on the research assistant highlights how it can help users in the following ways:
Accessibility and Usability:
- Democratises research access: Removes barriers for students lacking advanced database searching skills.
- Reduces research overwhelm: Helps students navigate through thousands of potential sources more effectively.
- Provides citation exemplars: Demonstrates proper academic referencing and synthesis techniques.
- Supports diverse learners: Multilingual capabilities benefit international students and cross-cultural research.
Academic Development:
- Starting point for research: Offers a structured entry into academic discourse.
- Iterative research methodology: Encourages exploration of multiple perspectives through varied results.
- Quality assurance: Built-in filtering ensures access to peer-reviewed, credible sources.
Limitations and risks
However, (as with any AI powered tool) the research assistant is not without limitations and potential drawbacks. As a user, it is important to understand how it works and what it can and cannot do. Based on the study conducted by the Open University, as a user you should be aware of the following potential issues:
Practical Limitations:
- Expect access frustration: Only 14% of users found all recommended sources available as full-text, so you may hit paywalls or unavailable resources.
- Understand quality limitations: The analysis only uses abstracts, not full texts, so you're getting incomplete information.
- Prepare for technical issues: You may encounter "robotic" AI language and broken links that disrupt your research flow
- Be aware of bias: The tool may reinforce existing academic biases and overlook diverse or minority perspectives you need for comprehensive research
User Experience Challenges:
- Expect a learning curve: You may find it confusing to switch between asking questions in natural language and using traditional academic search methods.
- Plan for extra support needs: Limited access to suggested resources may mean more trips to the library help desk.
- Manage your expectations: Don't assume AI can replace the skills and critical thinking you need to develop as a university student.
Academic Integrity Concerns:
- Beware of plagiarism temptation: The well-written AI summaries with references may seem easy to copy directly into your assignments—don't do this.
- Question AI authority: Remember that AI presents information as fact, but you still need to critically evaluate sources yourself.
- Recognise oversimplification: Complex academic topics may be presented without the nuance you need for university-level work.
Developing Wider Research Skills:
- Don't let your research skills atrophy: Over-relying on AI tools may weaken your ability to conduct independent, critical research.
- Avoid reduced critical thinking: The tool may discourage you from seeking out conflicting evidence or developing nuanced arguments in your work.
- Resist academic shortcuts: The tool's efficiency might tempt you to skip thorough, independent research that's essential for deep learning.
Understanding AI technology and its limitations
The Large Language Models that underpin AI technology are not very good at admitting they cannot help any further so as a user it is important to develop a sound understanding of 'AI literacy' and be able to critique the output and results AI tools give in order to recognise when they are below par.
