R
Research Independently First
Gather credible sources and form your own initial understanding before AI enters the process.
What do I know, what evidence have I found, and what do I think?
The framework
AI can support the work. It should not become the learner.
RAFT helps you decide what must remain yours, where AI can add value, and how to verify and transform what follows.
Before RAFT
Know the goal before you use the tool.
What am I expected to learn?
What AI use is allowed?
What have I already tried?
Research · Augment · Fact-check · Transform
Four actions that keep the learning visible.
A
Augment Your Work
Use AI to surface gaps, counterarguments, questions, and connections—not to produce the finished answer.
What might I be missing, and what should I investigate next?
F
Fact-Check Everything
Verify every claim, citation, quotation, and calculation against trustworthy original sources.
What confirms this, and does the source support the exact claim?
T
Transform Into Your Own Work
Build the final product through your own reasoning, structure, evidence, examples, and voice.
Can I explain and defend every part without the AI output?
A worked example
From a broad assignment to a defensible final product.
Imagine an assignment that asks a student to analyze a workplace change, apply course theories, and recommend a response.
Build a real starting point.
Read the case, gather credible reporting and research, map the timeline, apply two course theories, and write a preliminary claim before asking AI for help.
Ask for gaps—not an answer.
“I have applied these two theories and reached this preliminary conclusion. What other relevant frameworks might I have missed? Do not analyze the case for me.”
Trace every useful lead.
Find the original sources. Confirm they exist, check whether they support the exact claim, and record what was verified, revised, rejected, or remains uncertain.
Compose from understanding.
Close the AI output. Build the final analysis from verified notes, course concepts, original reasoning, and relevant experience. Then explain and defend the result.
At every step
Three responsibilities surround RAFT.
These are not extra letters. They are the conditions that make the process trustworthy and learning-centered.
Transparency
Follow and disclose AI use according to the local expectations.
Authenticity
Maintain your voice and demonstrate your actual understanding.
Responsibility
Own every claim, source, choice, and submission.
The final defense test
Can you explain it, verify it, and stand behind it?
- I can explain every important idea in my own words.
- I can show where the evidence came from.
- I can identify what AI did and what I did.
- I would be comfortable describing the process to my instructor.