AI in Education: 4 Real World Uses

Article by Hengtee Lim | January 16, 2020

AI in education is a field of evolving research, development, and implementation. In elementary to high school education, we’re seeing the application of intelligent tutors and smart learning content. This technology analyzes results to discover weaknesses, sets new challenges, and offers advice.

In college, speech to text systems support student note taking and lecture summaries. Virtual assistants help students keep up with schedules and upcoming exams. AI technology to automate administrative tasks and grading is constantly developing. Enterprises are also using automated AI systems to train, teach, and keep staff updated in dynamic work environments.

In this article, we’ll look at AI in education through a number of current technologies. We’ll examine what they are, where they’re in use, and the AI data and machine learning models behind them.


Intelligent Tutoring Systems

What it is: Intelligent tutoring systems are an AI-powered version of individual lessons, where AI takes the place of human tutors. So, the ideal intelligent tutoring system is just like a human tutor. It recognizes student errors, teaches them the correct answers, and tailors future work to address needs and weaknesses. Implementation of intelligent tutoring means giving more students access to individualized learning systems and attention that might otherwise be unavailable.

How it works: Intelligent tutoring systems are powered by databases filled with data annotated specifically for modules within topics of study. This includes potential answers, mistakes, and statistical data related to student progression, difficulty level, and related topics. To start, students work through lessons and the system compares their answers to the database, correcting them as necessary. As the student gets deeper into their study, the system recommends follow-up work and individualized material based on progress. The more students use an intelligent tutoring system, the more the database grows, and with the help of human fine-tuning, it grows more accurate and attentive.

Where it is used: The most well-known example of intelligent tutoring is Squirrel AI, in China. Squirrel boasts 2000 learning centers in 200 cities, with plans to expand through 2020. Squirrel AI’s individual tutoring system curates lessons for each student in subjects including math, English, and physics. Human teachers monitor these lessons in real-time and are available to answer questions and provide deeper levels of support. In the US, a similar system by the name of ALEKS (Assessment and LEarning in Knowledge Spaces) uses adaptive learning to teach maths, science, and business through its intelligent tutoring system.


Speech to Text Note Taking

What it is: Speech to text technology allows for automatic note-taking in classrooms. For students, it can mean more time listening to a lecture than worrying about which notes are the most important. For teachers, it can mean an easy way to prepare lecture summaries at the conclusion of a class, or to capture spontaneous discussion arising from student questions.

How it works: Speech to text technology relies on datasets of annotated speech samples. These audio samples are often broken into smaller chunks (or phonemes) that allow for machine learning models to understand what they hear and transcribe the audio accurately. Speech to text systems work best when they have large datasets catering to a wide variety of possible vocal inputs. This includes differences in age, sex, and race.

Where it is used: This area of AI in education is growing steadily. Otter, a company who once focused on automated note taking for meetings, also offers services for education. Their application allows users to set keywords in advance of recording a class, making it easy to search for specific content. Kidsense creates AI solutions for education for kids, and they train their voice to text systems specifically on the nuance of children’s speech for taking notes, doing tests, and practicing vocabulary.


Chatbots and Virtual Assistants

What it is: Chatbots and virtual assistants work to support students and teachers by increasing accessibility to information through automated conversational systems. For universities, chatbots serve to answer student questions about assignment deadlines, class locations, and exam dates. Not only do students get their answers immediately, but university staff are also freed up from answering repetitive questions.

How it works: Chatbots are trained on a corpus of text data specific to their use. In our example above, the training text data is likely a question-answer and conversational corpus relating to common enquiries and statements. This includes not just answers to questions, but also the variety of ways a question is asked. Virtual assistants, on the other hand, require similar audio datasets with a conversational focus.

Where it is used: Beacon is a real world example of a university chatbot. It is used by Staffordshire University to give students round-the-clock access to university information. It boasts information on student timetables and answers to 400 frequently asked questions regarding campus facilities and support services. Cognii’s virtual learning assistant, on the other hand, offers students personalized tutoring conversations and feedback on answers to open-response questions.


Smart Content and Text Analysis

What it is: Smart Content is an umbrella term for technology that summarizes papers and books. The technology creates smaller study guides for textbooks, and generates study/revision materials from core texts or lectures.

How it works: To summarize texts, smart content often relies on text analysis and natural language processing models. Text summarization can be as simple as extracting keywords from a text and combining them, or more complicated, such as paraphrasing key sections of text to create original summaries. As a model learns core keywords and concepts, it can begin to develop revision material such as quizzes. This material is then fine-tuned through datasets based on the specific nature of the desired revision material.

Where it is used: Learning platforms are a common example of smart content. Quizlet Learn utilizes a machine learning system to help students study and prepare for tests. It does this by creating tailored study plans based on current knowledge levels and upcoming test dates. On the enterprise level, Volley says they’ve developed a platform to synthesize an enterprise’s core content for employee training. For example, a cyber-security system could be taught through a series of automated summaries, flashcards, and quizzes. The systems teaches, tests, and revises as necessary.


Data for Future Development

The field of education is growing, and with new technology developing every day, it’s an area that will likely transform over the next decade. Along with the examples above, data analytics for course construction and AI for automated administrative tasks are also areas seeing ongoing research and development. The examples above are not the only uses for AI in education. There’s lots of room for improvement and innovation, but at the heart of it is data.

If you’re looking to implement your own AI solution, or are building automated technology to enhance the future of education, high-quality training data is key to the quality of your product. This is especially true in education, where errors and bias can have a massive impact on a student’s potential future. For more information regarding safe data collection, secure data privacy, and high-quality data annotation, get in touch.

Enhance your training data
The Author
Hengtee Lim

Hengtee is a writer with the Lionbridge marketing team. An Australian who now calls Tokyo home, you will often find him crafting short stories in cafes and coffee shops around the city.


Sign up to our newsletter for fresh developments from the world of training data. Lionbridge brings you interviews with industry experts, dataset collections and more.