Introduction to the Use of AI in Research
This introductory guide offers practical support for the use of AI in research at Uppsala University. It is not a set of rules or formal guidelines, but rather a straightforward tool to help you get started, work more efficiently, and avoid the most common pitfalls when using AI in your research.
Quick Start – Checklist
- Use Copilot Chat via your UU login.
- Never share sensitive or confidential information while using any AI tool.
- Need more AI tools? Contact IT support.
- Always verify AI-generated content.
- Document your use of AI tools.
Questions or concerns? Contact University IT Services via the Support Portal, by emailing itsupport@uu.se, or read more at the central AI information page for employees at Uppsala University.
What Can I Do Today?
You may use Copilot Chat in your research, provided you log in through the university system. Copilot Chat is classified in the same way as our collaborative platforms in SharePoint.
A guide for logging into Copilot is available here
Examples of Use Cases:
- Structuring ideas and research questions.
- Summarising literature you have uploaded.
- Providing language support and improving clarity in texts.
- Writing code or pseudocode, creating test cases, and analysing simple data.
- Drafting project descriptions, summaries, and applications.
Remember to check whether your research is subject to local AI guidelines.
How Do I Proceed If I Need More Tools?
- First, check if the tool is already available through the university (see below or this list of AI tool examples - in Swedish only).
- If not, use the Support Portal for help with licensing, security classification, and procurement. More information is available on the applications and licences page.
Key Points to Remember
Data Protection: GDPR always applies when processing personal data. Sensitive data will require special safeguards and ethical review.
Copyright: Check the rules governing the materials you process and whether you have the right to process material with AI.
Quality: AI can hallucinate or fabricate information. Always check facts and references.
Responsibility: You are always responsible for the final result, even if AI has been used in the process of creation/revision.
Transparency: Makes notes on how and when you employ AI tools in your research.
Examples of Useful AI Tools for Research
Chat- and Writing Support
ChatGPT (OpenAI)
General tool for text, analysis, translation, and code. Plus/Pro/Edu users have access to Deep Research and other advanced features.
Microsoft Copilot Chat
General tool for text, analysis, translation, and code. Allows creation of agents for recurring tasks.
Claude (Anthropic)
Chat tool that handles very long documents and conversations. Good for coding.
Gemini (Google)
Text and multimodal (text, image, code). Integrated with Google services.
Perplexity AI
Chat-based search and writing assistant, always with source references.
Search- and Research Overview
Consensus
Searches research articles and summarises study findings.
Connected Papers
Visualises relationships between articles via co-citation.
Iris.ai
Identifies and maps relevant research articles.
Paperguide
AI support for literature synthesis and research overviews.
Literature- and PDF Analysis
Elicit
Automates literature reviews, screening, and data extraction.
SciSpace Copilot
Allows you to ask questions about PDF articles and explains methods/analysis.
Scholarcy
Summarises articles into key highlights.
NotebookLM (Google)
Creates a personal AI based on your documents/notes.
Handling Personal Data in AI Tools
Many generative AI tools require sharing text or documents. If these contain personal data—such as names, personal numbers, email addresses, interview quotes, or other sensitive information—there are legal and ethical risks.
Some tools can help identify and protect personal data before sharing with AI. Examples include:
Personal Data Detectors: Open libraries, such as Microsoft’s Presidio, can be integrated into research environments to analyse and anonymise text before use in generative AI tools.
Local Filters: Local intermediaries that clean or mask personal data before data is sent to external AI services.
Recommendation: Always check if your material contains personal data. Avoid sharing sensitive data with external AI services where you cannot control storage and processing.
%20(2)%20AI_forskning_kugghjul_3.png)
Want to Know More?
The links below provide further information on training, networks, and support regarding AI at the university:
AI at UU in Brief
- Keep it simple to start: use Copilot Chat with your UU login.
- Follow data protection and copyright regulations.
- Always verify your results.
- Share your methods and experiences with your research networks.
This introductory guide is a starting point for engaging with generative AI. The content will be developed over time in collaboration with the Research Support Division and other units with expertise in the field.
All material on AI for staff at Uppsala University is collected on the Staff Portal.
- Note that there is a university-wide guideline for AI in teaching, and that your department may have local guidelines governing the use of AI.
- There is also a public page collecting links to news and information about AI in research (for visitors and prospective employees and students).