By Sal Gerardo
“A student is not a container you have to fill, but a torch you have to light up.”
— Albert Einstein
Education faces significant challenges. Some of these stem from the pandemic; others arise because schools and educational systems have been slow to adapt to a rapidly changing world and evolving industries. If we ignore these problems, they will intensify and have serious consequences for learners and society.
For context: I am unusual in education circles. I served for many years on a New Jersey school board, and professionally I work in computing, focusing on artificial intelligence (AI). I am a strong supporter of modern education and deeply respect teachers. Yet the system is becoming harder to defend, and educators increasingly shoulder disproportionate burdens.
Even before COVID-19, education was reaching a crossroads. The pandemic accelerated the need for change. With looming teacher shortages and growing demands on classrooms, technology—especially AI—will play a pivotal role. I believe that within a few years, educators who do not make use of AI tools will be at a serious disadvantage.
There is one thing teachers do better than any technology: teach. Teachers connect with students emotionally, inspire curiosity, and guide learning through human relationships. But if we allow teachers to be overwhelmed by administrative tasks, repetitive assessments, and content creation, we risk limiting their ability to teach. That is where AI can help.
Modern AI is becoming remarkably capable. It can take on repetitive mechanical tasks, handle assessments, and manage personalized content delivery—freeing teachers to focus on human-centered activities like mentoring, classroom dynamics, and deep instruction. AI can streamline the minutiae so teachers can teach.
How AI systems learn and act
Developing AI that supports education requires integrating multiple capabilities. Tasks like accurate translation or meaningful dialogue are “AI-complete” because they demand natural language understanding, reasoning, domain knowledge, and an appreciation of intent and context. Unlike the human brain, which stores and learns meanings concurrently through sensory experience, current computational systems separate storage and learning. Effective educational AI combines three complementary components:
- The Knowledge Library — an organized repository of domain knowledge and teaching expertise.
- The Intelligent Grammar — mechanisms that represent meaning, relationships, and discourse structure.
- The Learning Process — adaptive algorithms that infer concepts, track progress, and update the system’s knowledge.
When these parts work together, AI can support large-scale, meaningful learning interactions and scale expertise globally.
Three core technologies in advanced educational AI
-
Learning and Meaning Engine (LME)
- Organizes human knowledge into structured libraries.
- Learns from conversations, documents, websites, and databases.
- Extracts concepts, narratives, and expert behaviors using machine learning and structured induction.
- Models discourse and meaning with controlled constraints to support reliable tutoring.
-
Artificial Intelligence Behavior
- Performs data engineering and executes goal-directed behaviors.
- Adapts dynamically to meet objectives and tracks success continuously.
- Maintains an extensible Knowledge Library with expert behaviors, including analytics and STEM tutoring.
- Can generate software behaviors and instructional sequences from stored expertise.
-
Natural Language Avatar
- Engages learners through natural-language dialogue, including Socratic questioning.
- Supports open Q&A and models interlocutor behavior through conversation.
- Monitors sentiment and comprehension, adjusting responses to build rapport and support learning.
Socratic tutoring and the skills it requires
Socratic tutoring centers on asking thoughtful questions to promote reasoning and mastery. Implementing it computationally requires a procedural model that coordinates multiple complex behaviors. Based on research, effective Socratic tutoring involves a set of core skills, including:
- Speaking and listening, with translation when necessary, for teacher-student exchanges.
- Reading and writing, and processing student- and teacher-generated text.
- Formulating appropriate question-and-answer formats for different learning tasks.
- Learning and retaining subject matter to be taught.
- Learning and remembering information about individual students.
- Understanding and inferring meaning in student responses.
- Planning teacher utterances and instructional sequences.
- Conducting sustained natural-language dialogue.
- Self-organizing subject-matter expertise about Socratic methods.
- Maintaining goal-directed behavior focused on student mastery.
- Reasoning about student answers and misconceptions.
- Assessing mastery through dialog and open-ended interaction.
- Adapting the teaching process to improve student outcomes.
Research consistently shows learners gain more from one-to-one tutoring tailored to their pace and questions. Advanced AI tutors will conduct natural-language Socratic conversations, continuously monitor comprehension, and patiently guide learners until true understanding is achieved. This dialog-based assessment is stronger than multiple-choice testing, which can be gamed without genuine learning.
Emerging AI solutions are being built around meaning engines that simulate human ways of understanding words, sentences, and stories. Unlike scripted chatbots or simple Q&A systems, these systems aim to generate unscripted, context-sensitive dialogue by drawing on a growing Knowledge Library. As the library expands, the AI can converse across domains and personalize instruction for each student.
How AI will augment teachers
AI will serve two complementary roles in classrooms:
- Provide a Socratic tutor for every student under teacher supervision
- Use dialog-based tutoring to deliver personalized content.
- Assess learning style and level, and respond to students’ emotions.
- Provide context-aware answers and remedial or enrichment support.
- Act as an assistive partner for teachers
- Collaborate with teachers and learn their preferences.
- Automatically generate course materials from textbooks, teacher inputs, and vetted online sources.
- Offer real-time assessment and alert teachers to barriers to learning.
Key differences between current “AI tutors” and next-generation systems include genuine natural-language Socratic interaction, real-time monitoring of individual learning, automatic creation of course content, and scalable architectures that can handle accelerating STEM knowledge and instructional needs.
Imagine shifting from a system that expects teachers to spend many hours each week on lesson creation and administrative tasks to one where AI handles the heavy lifting. Teachers would regain time to perform the human work: mentoring, inspiring, and engaging students in deep thinking. AI won’t replace teachers; it will free them to be more effective and more relational.
Teachers are dedicated professionals. With thoughtful deployment, AI can remove repetitive duties, provide accurate assessments, and enable teachers to focus on the critical human aspects of education. In doing so, teachers will remain the torchbearers who ignite a generation’s curiosity and potential.
About the Author
Sal Gerardo is chairman of Global Education Media.
This article was first published in (ET) Magazine.