AI: Will it Replace You, or Empower You? A Teacher’s Practical Journey of Human-Machine Collaboration

I understand many educators fear being replaced by generative AI, but in my view, things should be considered from two perspectives:

1. What cannot be changed: The trend of generative AI is unstoppable; it will eventually become widespread and a native skill for the next generation.
2. What can be changed: How we can collaborate with it to co-create greater value.

Today, I want to share the amazing transformations that will occur in your work model and personal value when you start thinking about how to “collaborate” with AI, rather than letting it “replace” you.

The Evolution of a Teaching Process: From “Purely Experience-Based” to “Human-AI Collaboration”

Since I developed a numerical analysis model for GMAT scores, while the study plans for students have become more detailed, it has also increased my working hours. Before a 1-1.5 hour consultation with a student, I often need to do an additional hour of preliminary analysis.

Past: Time-Consuming and Error-Prone “Manual Analysis”

Before integrating AI, my process was: manually reviewing student score report Excel files, coloring cells one by one, and taking notes as I went. This not only resulted in disorganized formatting but also risked missing important analytical steps due to human oversight. This entire process took an hour just to analyze one student’s score report.

Present: 60% More Efficient “AI-Assisted Process”

Yesterday, I had a sudden idea: I communicated my analysis logic to AI, and unexpectedly, it actually wrote the analysis tool as a VBA macro! The current process is:

1. One-click analysis: Open the score report Excel file and run the macro. AI automatically highlights areas I need to focus on (e.g., too fast, too slow, simple questions of the same type answered incorrectly).
2. SOP guidance: I open the “Personalized Analysis SOP” that AI organized for me. While reviewing this SOP, I analyze the score report, ensuring it’s structured and nothing is missed.
3. Voice notes and automatic organization: During the analysis, I record the entire process. AI automatically transcribes the audio and outputs structured analysis suggestions according to my predefined template format. AI even proactively reminds me of areas I might have overlooked based on my own analysis logic.

The results are:

• The time I spent on repetitive tasks was reduced by 60%.
• All analyses remain complete, never missing key points, and the content is 100% based on my personal experience and expertise.
• The output format is highly organized, making it easy to explain to students and convenient for them to review later.

How I “Trained” AI to Learn My Analysis Logic

All these changes stem from a shift in a core idea: I don’t demand fully automatic analysis from AI; instead, I think about “How can AI assist me in analysis?” This led to the following training process:

1. Provide samples: I analyze while recording my thought process, then submit both the audio files and analysis notes to AI as multiple samples.
2. Induce logic: I ask AI to compare these samples to identify the aspects, characteristics, and SOP of my analysis, effectively “formatting” and “standardizing” my implicit knowledge.
3. Develop tools: Finally, based on this standardized analysis method, I have AI automatically write VBA macros to complete this highly efficient auxiliary tool.

Honestly, without AI, I might not have been able to understand the numerous statistical papers on GMAT official algorithms; even if I did, I wouldn’t have been able to write the algorithms into Python scripts; and I certainly wouldn’t have been able to develop such an efficient analysis method. These were all unimaginable to me before the advent of AI.

My Conclusion: AI is a “Kingmaker,” Not a Replacer

In the past, I felt I had hit a bottleneck in my teaching career, constantly answering the same questions and handling daily chores. With generative AI, it not only helps students solve many basic problems but also frees me from time-consuming, repetitive tasks, allowing me more time to focus on teaching research and development.

I haven’t lost my job, nor have my student numbers decreased—despite being the one most actively teaching students how to use AI.

• Thirty years ago, when internet technology first emerged, many believed “computers would replace humans.” The truth is: people who could use computers replaced those who couldn’t.
• Five years ago, when AlphaGo burst onto the scene, some also thought chess players would become obsolete. The truth is: all top chess players now use AI for training, which has elevated the competitive level of Go.

So, in my view, AI is a “Kingmaker.” If you have the potential to be a king, it can make you one; if you are mediocre or even resist collaborating with it, you miss the best opportunity of this era.

Similarly, as a mechanic, you can be an experienced old master who relies purely on intuition, or you can conceptualize, standardize, and create SOPs for your experience, reducing unnecessary effort in your life and making your knowledge and skills easier to disseminate and pass on.

These are some thoughts I wanted to share with everyone.

Related posts:

Cracking GMAT Reading Comprehension: An AI Tool for an 80-Point Score Boost GMAT Study Review with Voice Recognition AI Don’t Treat AI as a Slave Anymore! Teach You the Four-Step Mindset for “Human-Machine Collaboration” GMAT Prep: GPT’s Precision Questions & Deep Dives