August 6, 2025

Managing AI in Research

What Every Research Administrator Needs to Know

Your researchers are already using AI tools like ChatGPT and Claude. The question isn't whether artificial intelligence will transform research at your institution—it's whether you'll be ready to manage that transformation effectively.

The Reality Check

Large Language Models (LLMs) aren't just productivity tools anymore. Researchers are using them to:

• Analyze thousands of documents in hours instead of months
• Generate synthetic survey data when human participants aren't available
• Code qualitative interviews at unprecedented scale
• Extract insights from vast collections of government records and social media

This isn't coming—it's here. And it requires institutional oversight.

The TaMPER Framework: Your 5-Point Action Plan

A new framework by the AI4RA team at university of Idaho called TaMPER gives research administrators a clear roadmap for managing AI in research. Think of it as quality control for the AI age:

T - Task: What is the AI actually doing?

• Require researchers to clearly document AI use cases
• Distinguish legitimate research applications from general productivity use

M - Model: Which AI system and how is it configured?

• Address data privacy (commercial AI = data leaves your institution)
• Manage costs and technical infrastructure needs
• Consider local hosting for sensitive research

P - Prompt: How are researchers interacting with the AI?

• Poor prompts = biased results
• Require documentation of all AI instructions used

E - Evaluation: How do we verify the AI did a good job?

• Mandate validation protocols for AI-assisted research
• Set minimum standards for quality checks

R - Reporting: What must be disclosed about AI use?

• Ensure transparency for peer review and replication
• Meet emerging journal and funder requirements

Three Immediate Actions You Can Take

1. Survey Your Situation

Find out what's already happening. Many researchers are using AI tools without institutional guidance—creating potential compliance and quality issues.

2. Develop Clear Policies

Create guidelines that distinguish between:

• Low-risk uses (analyzing public documents)
• High-risk applications (sensitive data, human subjects research)
• Prohibited uses (confidential data on commercial platforms)

3. Build Support Infrastructure

Provide training and consultation services. Researchers need both technical skills and methodological guidance to use AI responsibly.

Why This Matters Now

Funding agencies are beginning to require AI use disclosure in proposals and reports.

Journals are establishing new standards for AI transparency in publications.

Institutional review boards may need updated protocols for AI-assisted research.

Legal and compliance teams need clear policies for data privacy and risk management.

The Bottom Line

AI in research is like the introduction of statistical software in the 1980s or online databases in the 1990s—a fundamental shift that requires institutional adaptation. The difference is the speed of change and the breadth of impact.

Your researchers will use these tools. Your choice is whether they'll use them well, with proper oversight, training, and quality controls—or whether they'll improvise without institutional support.

The TaMPER framework gives you a structured way to ensure AI enhances rather than undermines research quality at your institution.

Ready to dig deeper? Check out our comprehensive guide to implementing AI governance in research institutions.