Recently, an MIT professor used Google Notebook LM to demonstrated something remarkable.
Grading 47 essays usually means one thing:
A long night.
Cold coffee.
And mental exhaustion.
But recently a MIT Professor graded 47 essays in about 12–15 minutes.
Not by rushing.
Not by cutting corners.
But by using Google’s NotebookLM strategically.
This isn’t about replacing teachers.
It’s about removing repetitive workload so educators can focus on real teaching.
Let’s break down exactly how he did it — step by step.
Step 1: Create a Dedicated Notebook
He started by going to:
notebooklm.google.com
It’s free with a Google account.
He created a new notebook and named it clearly:
“Essay Grading – Class XYZ”
Why this matters:
NotebookLM works best when everything lives inside one organized space.
It reads only what you upload.
That keeps outputs focused and accurate.
Step 2: Upload Everything at Once
This is where most people miss the power of NotebookLM.
He didn’t upload just the essays.
He uploaded everything relevant:
• All 47 student essays (PDF or Word)
• The grading rubric (with scoring criteria)
• Course materials (lectures, slides, readings)
• Previous submissions from the same students
Why upload all of this?
Because AI performs best when context is complete.
Instead of giving vague grading suggestions, NotebookLM had:
• The expectations
• The reference material
• The student writing history
In other words, it had the full academic environment.
Step 3: First Prompt — Grade Against the Rubric
He typed a very clear instruction:
“Evaluate each paper against the rubric I uploaded. Give a score for every criterion and flag any papers that look off or don’t follow expected patterns.”
That’s it.
NotebookLM then:
• Read all 47 essays
• Matched them against the rubric
• Assigned scores for each grading criterion
• Added short notes explaining the reasoning
And it did this in seconds.
Not summaries.
Not guesses.
Structured grading aligned with his rubric.
Important detail:
He didn’t ask it to “grade essays.”
He asked it to evaluate against a rubric.
Specific instruction = specific output.
Step 4: Second Prompt — Detect Suspicious Writing Patterns
Next, he focused on integrity.
He typed:
“Now compare the writing style of each essay to the previous submissions from the same students. Highlight any big style changes that might mean they copied or used AI.”
This is powerful.
Instead of relying on traditional plagiarism tools, he asked the AI to:
• Compare tone
• Compare vocabulary complexity
• Compare sentence structure
• Compare writing maturity
NotebookLM identified three suspicious cases he would likely have missed.
Not definitive accusations.
But flags worth reviewing.
This step saved him hours of manual comparison.
Step 5: Final Prompt — Generate Personalized Feedback
Grading numbers is easy.
Personalized feedback is what takes time.
His final prompt:
“For each student, write short personalized feedback. Point out their weak spots and tell them exactly which course materials (lectures/readings) they should review to improve.”
This changed everything.
Because he had already uploaded course materials, NotebookLM could:
• Reference specific lectures
• Suggest particular readings
• Connect weaknesses to learning resources
Instead of generic comments like:
“Needs improvement.”
Students received feedback like:
“Your argument lacked evidence. Review Lecture 4 on evidence integration and revisit the assigned reading on structured argumentation.”
That’s meaningful feedback.
And it was generated instantly.
Why Was It So Fast?
Three key reasons.
1. Everything Was in One Place
NotebookLM works best when all relevant documents are uploaded.
It didn’t need to “search the internet.”
It didn’t need external tools.
All essays, rubric, and course materials were inside one controlled environment.
2. AI Reads at Machine Speed
Humans read 200–300 words per minute.
AI processes thousands of pages almost instantly.
That comparison is not about superiority.
It’s about different capabilities.
Let AI handle pattern recognition.
Let humans handle judgment.
3. He Asked Only Three Clear Questions
Instead of dozens of back-and-forth instructions, he asked:
- Grade against the rubric.
- Compare writing style.
- Generate personalized feedback.
Clear tasks.
Clear structure.
Clear output.
What This Means for Teachers
This is not about replacing educators.
It’s about reclaiming time.
Imagine what teachers could do with 10 extra hours:
• Improve lesson quality
• Mentor students individually
• Create better assessments
• Focus on struggling learners
AI becomes an assistant.
Not a substitute.
Can Non-Tech Educators Use This?
Yes.
No coding required.
No technical background required.
The real skill here is not programming.
It’s structured thinking.
You need:
• A clear rubric
• Organized materials
• Specific prompts
That’s it.
If you can upload files and type a question, you can use this workflow.
Important Ethical Considerations
Before adopting AI grading, institutions should think about:
• Transparency with students
• Data privacy policies
• Final human review
• Academic fairness
AI should assist decisions, not finalize them independently.
The professor still reviewed the outputs.
He didn’t blindly accept them.
That balance matters.
Broader Implication: AI as Workflow Multiplier
This example isn’t just about grading.
It shows something bigger.
When AI has structured context and clear instructions, it becomes a workflow multiplier.
This same approach can be applied to:
• HR reviewing resumes
• Managers evaluating reports
• Researchers analyzing documents
• Lawyers reviewing case files
The pattern remains the same:
Upload context.
Ask structured questions.
Review intelligently.
The Real Lesson
The real lesson isn’t “AI grades essays.”
The real lesson is this:
Professionals who know how to structure AI workflows will move faster than those who don’t.
The advantage is not technical skill.
It’s clarity of thinking.
Should Every Teacher Use This?
Not blindly.
But ignoring it may not be wise either.
Students are already using AI tools.
Institutions are experimenting with AI systems.
The better approach is:
Learn it.
Test it.
Use it responsibly.
Because the future of education won’t be:
Teacher vs AI.
It will be:
Teacher + AI.
Final Thought
47 essays.
15 minutes.
That sounds futuristic.
But the real story isn’t speed.
It’s leverage.
AI does not eliminate effort.
It amplifies structured effort.
And that changes everything.
What to Read Next
• 12 AI Skills Non-Tech Professionals Should Learn
• 10 Smart Ways to Use Perplexity for Daily Productivity & Learning
• Elon Musk’s 30–36 Month AI Prediction
Notion Pack
AI for Teachers Workflow Pack Download here