Proof of concept for socratic ai
The first instinct of any teacher at a high school or college level is to tell their students that any essays or essay question responses must be original text, with no assistance from generative AI. Personally, I think this may be misguided.
Generative AI is likely to improve over the course of the following years, and students leaving college will be expected to use it. Banning the use of generative AI in 2025 is like banning the use of calculators in math class in 1985. Pointless. So let's take another look at what it means to use generative AI in a classroom setting.
The goal of teaching is for students to learn key concepts that will allow them to think creatively at a high level in the future. Solving an example math problem is not about being able to solve that particular problem 20 years later. It's about knowing how to apply concepts from math to real life, so that when real life throws a novel problem at an adult, the skills they learned in school can be used to solve that novel problem more quickly and accurately. Learning is not about solving a particular problem. It's about practice and getting to an accurate answer quickly.
For centuries, teachers and professors, wanting evidence that their students have learned specific concepts, have assigned essays for them to write. Not because the teacher needs to read another essay. The essays are Key Indicators of learning. By reading an essay, a teacher can evaluate whether a student has learned the expected material.
In the age of generative AI, this changes. The essay is no longer a key indicator of mastery. Should we abandon essays? Perhaps. Or perhaps we should take a look at how teachers evaluated student learning before the practice of assigning essays was common. I'm going all the way back to Socrates.
The Socratic method is a form of teaching and dialogue based on asking and answering questions to stimulate critical thinking and draw out underlying assumptions. Named after the Greek philosopher Socrates, it involves a cooperative, argumentative dialogue where participants are encouraged to examine their beliefs through a series of probing questions. The goal is not necessarily to arrive at a final answer, but to deepen understanding and expose contradictions in thought.
I propose a plug in for LLM systems like Moodle and Canvas where a student submits an essay and the plug in evaluates, using GPTZero, if the paper uses AI, and regardless, asks the student a series of challenging questions on the topic of the paper using the socratic method. The goal is not to determine if the student used AI to complete the assignment, although that is a factor in the evaluation. The goal is to determine if the student actually learned the material.
This project is a POC of this idea