Insight

AI in education, Use Cases and Applications

Expectations and concerns about closing the training gap and personalizing training

2023

.

09

.

12

by

Sangsun Moon

AI in education, Use Cases and Applications

Bill Gates said, "Artificial Intelligence (AI) technology will drive education innovation." In addition to addressing educational inequality, the future of technology will also have an impact on reducing the cost of private education, which will go a long way toward improving the academic achievement of low-income students in the U.S. and around the world, especially as AI becomes more personalized to students, allowing for instant feedback.

While there is a lot of interest in AI-powered education, there are also a lot of concerns about the effectiveness of education without human feedback. In this article, we'll take a look at what's most important for the future of AI education, and how we can learn from our hopes and concerns.

‍

What are the 5 AI Trend keywords for 2023?

‍

What is AI-powered education?

Definition

‍

10 Ways AI in Education is Transforming the Industry

‍

AI-powered education is a personalized program that uses AI to analyze learning data, including what students know and don't know, and what subjects or areas they have strengths in, to diagnose them and design a way for them to learn. It can be used as a teacher's assistant for personalized instruction, or it can be used in online platforms for learning analytics. Naturally, successful data analysis for personalized education requires sufficiently diverse learning time and log data.

In the traditional way of education, learning takes place through the interaction between the teacher and the student. This is a process in which the teacher learns the characteristics of the learner and leads the education based on them. The traditional way is a structure in which the teacher leads the learner, creating a dependent relationship. The subjective judgment of the teacher was important in determining the educational policy for the student, and it was difficult for the student to escape the dependent structure in the one-dimensional teacher-student relationship. However, when AI is applied to education, the advantage is that the education can be more objectively tailored to the learner's level by digitalizing the learner's educational habits and patterns, and the education can be continued consistently even if the teacher changes.

‍

An elementary school in the US with 25 students and 25 AI teachers. — An elementary school in the United States with 25 students and 25 AI teachers. (Source: Chosun Ilbo)

‍

Recently, AI-powered education programs and systems have gained a lot of traction and market attention. By accumulating data on students, they can categorize their strengths and weaknesses, and personalize learning by adjusting difficulty levels and repetitive learning curricula.

‍

Why AI-powered education matters

Personalized education

‍

The CEO and founder of Khan Academy, Sai Kahn recently spoke out about the importance of AI education in a TED Talk. "We are on the cusp of leveraging AI for the most positive change we've ever seen in education," he declares of current education trends. We can give every student on the planet an amazing personal tutor in the form of AI, and every teacher can have an amazing teaching assistant in the form of AI, he added.

According to a report by McKinsey, the biggest challenge to personalized education is the lack of physical time for teachers. With so many students and limited time for teachers, personalized education would probably remain a pipe dream without the help of technology. However, AI-powered virtual teachers can provide round-the-clock support to students. Personalized learning solutions can also help students with special needs reach their full potential. In addition, immersive learning content built on learning data can make education a more engaging and memorable experience.

‍

Increase teacher productivity

At a recent AI education conference, Daniel Schwartz, dean of the Graduate School of Education, explained that "technology offers the possibility of universal access to fundamentally new ways of teaching." In fact, many teachers are already familiar with EdTech, which combines education and technology. Code.org's Hadi Partovi, 30 percent of teachers are now using AI to actually develop lesson plans.

Many teachers have already realized that all of the routine and simple tasks in the classroom can be Chatbots are delegating to intelligent agents. Teachers are free to focus on other, more in-depth tasks related to quality education. With the personalized learning solutions provided by AI-powered education programs, they are able to build more effective curricula.

‍

Overcoming educational alienation

AI technology can extend educational opportunities to students in geographically and economically underserved areas or social groups. Online courses and eLearning platforms are accessible from anywhere and can be used by anyone with an internet connection. These AI-enabled digital educational resources provide access and equality to students in geographically remote or hard-to-reach areas.

In addition, AI technology has been shown to be a key factor in how Multilingual processing and automatic translation help bridge language barriers. AI systems that communicate and understand content in multiple languages can make content more accessible and understandable to students studying in foreign languages. They can also mitigate cultural discrimination and marginalization by developing contextualized educational materials and content that reflect different cultures and customs.

‍

Concerns about educational AI

The realities of digital education

‍

Digital textbooks — Paper textbooks are going away... K-12 to use 'digital textbooks' from 2025 (Source: )

‍

Digital education is a rapidly developing field in the modern world, but its adoption and application comes with a number of practical challenges.

Reading comprehension researchers claim that we get better comprehension and deeper thinking when we read on paper. This is why digital textbooks have been around for quite some time, but some experts are still skeptical of the technology. There are concerns that it could take away from young students the benefits of reading on paper.

The question is, are regular schools sufficiently prepared for digital education? Forcing new pedagogical approaches without adequate teacher training, infrastructure, and consensus on the adoption of digital education can backfire.

There are many counterarguments to the idea that personalized education can help students who are falling behind. For starters, there's the practical challenge of how to personalize instruction when classrooms are still so overcrowded. Even if AI can personalize learning, it's still not a substitute for a teacher to know and provide feedback to each student. There's also the concern that different home environments can lead to different levels of digital literacy and readiness to learn, which could widen the educational gap.

‍

The reality of personalized education

‍

Personalized learning and interaction — An environment where you learn and grow by interacting with people is hard to replace (Image source: The Hechinger Report)

‍

Artificial Intelligence use data and algorithm, which works on the basis of Lack of real-world interaction. There is a lack of Human interaction plays an important role in the educational process. In this regard, it can be difficult for an AI system alone to provide sufficient support and motivation to students.

AI systems also require large amounts of data. However, this data collection and processing process can introduce bias. For example, data may be biased based on preferred learning styles or cultural differences, which can raise issues around diversity and inclusion. Social stewardship and ethical concerns around the use of AI systems are also important. It is necessary to ensure transparency and fairness about the decision-making power of AI systems, how they are evaluated, and any algorithmic biases.

There are also concerns about the adverse effects of 'personalized education' on socialization. One of the important goals of digital and AI-powered education is to provide a personalized learning experience. But the reason we go to school is because it also serves the important purpose of socialization for children who are still young. How will personalized education change the way children learn and interact together? If the learning process is going to be vastly different for each child, students in similar courses may be able to interact with each other based on their understanding of the content, while students separated by different courses may feel isolated, in the same classroom but in a different world.

‍

Educational AI Use Cases - AI Platform

Khan Academy - Khanmigo

‍

Khanmigo Education AI Guide | Khan Academy

‍

Large Language Model (LLM) was developed, it was the education industry that was most disrupted. Concerns about cheating by utilizing artificial intelligence technology that has more than encyclopedic knowledge and answers like a human to take the place of tests and essay assignments have led some schools in the United States to ChatGPT was even blocked.

Schools around the world are now using Generative AI in education, we're working to find the best practices and guidelines for its use. We want to maximize the effectiveness of personalized learning with generative AI, but we also want to ensure that We need a way to avoid inaccurate or biased errors. He also explained that there is a need to ensure that an over-reliance on technology does not lead to side effects such as diminished learning ability or bipolarity.

In response, Khan Acamemy, a non-profit educational organization, introduced Khanmigo, a GPT-4-based AI tutor. Khanmigo is somewhat different from the traditional ChatGPT that provides instant answers. It focuses on having more sophisticated conversations based on GPT-4, acting as a guide to support teachers and students in their learning activities.

In a pilot video released by Khan Academy, a student asking a math question is shown suggesting the thinking and learning required to solve the problem. If commercialized after continued pilot testing, it would mark the first time AI has been applied to the world's largest free online education service used by tens of millions of people.

‍

TagHive - Class Saathi

South Korean edtech company Taghive has been in the news for delivering its smart classroom product, Class Saathi, to schools in India. TagHive said that it has supplied the product to 2,080 schools in 52 districts of Madhya Pradesh, India's second-largest state, since 2020, with a second round of exports taking place in stages, establishing a successful collaborative relationship. It was quite a big deal for a homegrown edtech startup to be recognized by the Indian public education sector.

‍

Class Sati is designed to allow students to participate in class by having a device with an OX and a number on it, and clicking on answers to quizzes given by the teacher. Through the smartphone app, the teacher can see, collect, and evaluate the opinions of the entire class in real time. This allows for effective classroom management, including adjusting the level of instruction based on individual student engagement.

Taghavi explained that even in schools in parts of India that do not have smart education infrastructure, Class Sati can be used to achieve high educational impact at a low cost. If you want to see Class Sati in action in India, check out this video.

‍

QANDA

‍

Kwanda is a learning app that lets you take a picture of a problem you don't understand, upload it, and AI reads it to provide personalized solutions and learning content. Recently, Kwanda's operator, Espresso, announced the beta launch of Polly, a problem-solving tutor based on generative AI.

Poly has developed its own OCR (optical character reading) technology, which he explained was implemented as if a real teacher were sitting next to him. Instead of just showing the answer and the solution at once, it explains the solution step-by-step to help students understand the context. By leveraging its 5.8 billion pieces of education domain-specific data and AI technology, Maspresso has been able to create an LLM-based AI Tutor service, which the company says it will continue to enhance.

The reason why the education industry has been hesitant to adopt generative AI is because of the problem of providing accurate information. The biggest flaw in the current GPT-4 is the illusion, which is the error of answering incorrect information as if it were real. To overcome this, Maspresso explained that it will optimize the LLM and dramatically improve the correct answer rate by using proprietary technologies and data, such as OCR technology that accurately recognizes shapes and graphs, and algorithms for calculating formulas.

‍

What is GAN, the core technology of Generative AI?

‍

Duolingo

‍

Duolingo's AI analyzes user responses to identify common grammar mistakes and offers corrections and suggestions. As learners write sentences, AI analyzes their content to identify grammatical errors and help them correct them with correct wording. We've also implemented an autosuggest feature in the Stories free-form writing exercise, where AI automatically gives suggestions and advice while learners are writing on a given topic. This allows students to focus on using language more naturally and accurately, while also improving their writing skills.

It also provides personalized feedback on pronunciation by analyzing the sounds and patterns in your voice. As learners pronounce a word or sentence, the AI evaluates the content and offers individualized pronunciation corrections and advice, giving them direction to improve their voice and weaknesses.

Since 2021, Duolingo has been successfully utilizing GPT-3 in its proficiency assessment system, which was recently upgraded to GPT-4. With this, Duolingo has begun experimenting with new item types to make the proficiency assessment system more engaging and effective.

‍

Technology in Educational AI

From the examples above, we can see that despite the many concerns about AI in education, the technology is being utilized in many different directions. Here are some of the AI technologies that have made EdTech possible today.

‍

Natural Language Processing (NLP)
Technology that helps computers understand and process human language
Interact with learners and Use as a personalized learning tool, like a virtual assistant that answers questions or provides feedback automatically.
Machine Learning
Data in computer system, technology that has the ability to learn patterns and make predictions
Used to analyze students' learning patterns and performance to provide individualized and personalized educational paths, or to automatically perform grading and feedback.
Computer Vision
Computer vision is the ability of a computer to understand and interpret image or video data
In the field of education, computer vision is used for facial recognition, Used for object detection and automatic annotation, etc.
Reinforcement Learning
A technique in which an agent interacts with the environment and self-learns behaviors to maximize rewards
Used to teach students problem-solving skills in a virtual simulation environment.

‍

Conclusion: It's important to design a BM that can leverage educational AI

Personalized education AI enables the education industry to provide personalized learning experiences and tailor instruction to students' abilities and needs. This is a way to mitigate educational exclusion by identifying each student's strengths and weaknesses and providing them with the optimal learning path.

In addition, educational AI can enable effective learning assessment with automated assessment and feedback systems. AI technology's ability to analyze large amounts of data and recognize patterns enables early prediction and intervention, providing timely help to struggling students.

The use of AI in education also holds great potential for online courses and eLearning platforms. Digital educational resources that can be accessed from anywhere, regardless of geographical or economic constraints, can make a huge difference in terms of accessibility and equality. Additionally, cultural discrimination and language barriers can be addressed with a variety of language processing and automatic translation capabilities.

However, we must also recognize some of the concerns that come with the adoption of AI. For example, the lack of human interaction or data dependency. These are concerns that need to be addressed with the right complements and ethical considerations.

Therefore, we need to devise a strategy and direction for the adoption of educational AI in the education industry. Already, many teachers believe that AI should be used as a research tool, so students will see the output of generative AI as a starting point, not a finished product. The education industry, as well as the AI industry, needs to reframe how we use AI to support the 3Cs of technology engagement while fostering critical thinking. Students will find it easier and more enjoyable to organize, collaborate, and create with educational AI.

In conclusion, the adoption of AI in education is not without its concerns, but the potential is significant. We need to focus on providing individualized learning experiences and equal educational opportunities through customized educational AI and envisioning different ways to use it, and we need to collaborate among relevant actors to ensure continuous progress and create a more enriching classroom environment in the future.

‍

AI in education, Use Cases and Applications

Expectations and concerns about closing the training gap and personalizing training