Three Roles AI Can Play
Dominik Lukes identifies three distinct modalities for interacting with AI in a research context. Understanding which role you need at any given moment is key to effective usage.
The tireless administrative aid
As an Assistant, AI helps clear the clutter. It takes on the repetitive, cognitive drudgery that slows research down. It doesn't necessarily generate new knowledge, but it organizes existing workflows.
- Drafting emails to participants or collaborators
- Formatting bibliographies and citations
- Summarizing meeting notes into action items
- Scheduling and project management
"Here are my rough notes from the lab meeting. Please summarize the key decisions made regarding the sample size and create a to-do list for the graduate students."
The pattern seeker
As an Analyst, AI acts as a sounding board and a second pair of eyes. It can process vast amounts of text to find themes, contradictions, or gaps that a human might miss due to fatigue.
- Coding qualitative data (interviews, survey responses)
- Identifying logical fallacies in arguments
- Suggesting statistical approaches for datasets
- Comparing and contrasting theoretical frameworks
"I am pasting three abstracts from recent papers on cognitive load theory. Identify the conflicting definitions of 'intrinsic load' across these three texts."
The technical enabler
The most transformative role. You no longer need to be a master programmer to create custom software. AI can write the code you need to solve specific, one-off research problems.
- Writing Python scripts to scrape public data
- Creating Regex patterns for text cleaning
- Generating R code for specific visualizations
- Debugging Excel formulas or SQL queries
"Write a Python script using Beautiful Soup to extract all table data from a list of URLs in a CSV file and save the results as a new JSON file."