How is AI changing education

This image collected from from Wikimedia. No changes were made to this image.License inforation

AI implementations will transform the teacher student paradigm and change how teachers teach and how students learn. The objective of this page is to show you real world examples of how AI is changing education. Through the examples we will demonstrate current technologies being applied to learning.

Personalized Instruction

Personalizing instruction to meet individual student needs has been a priority for educators for years. Unfortunately, this task is time consuming and requires observational practices which are not possible within the construct realities of educational systems. By constructing learning content and LMS systems based on AI predictive models, AI can begin to play a larger role in detecting learning needs and responding with appropriate content. The following companies have implemented systems with these capabilities.

  • Knewton Alta
    • Alta is a responsive course delivery platform which uses AI technology to guide students towards mastery by analyzing learning deficits and presenting learners with instructional materials and remediated content as needed. This technology utilizes a just in time delivery model to ensure students are supported when they need support.
This video is intended as an extension activity. Viewing the video is not required.
  • Century Tech
    • Century tech offers an AI supported LMS system which uses student input to customize learning ‘pathways’ in English, math, and sciences. The Century system is geared for grades 3 – 6. The system also provides teachers with access to specific learning resources, learning analytic information, and instructional guidance. Not only does the system provide individualized content but also provides feedback to support learning.
This video is intended as an extension activity. Viewing the video is not required.

Other personalized learning platforms supported by AI: Carnegie Learning, Kidaptive

Natural Language Processing

Natural language processing (NLP), in the context of AI, is the ability for programs to analyze grammatical structures of sentences and the meaning of individual words in order to deliver an output. In education this can look like text to speech, speech to text, grammar and spell checking, translation tools, or chat bot technologies. The following companies use NLP technologies to support education.

  • KidSense
    • KidSense has created an NLP product designed to specifically transfer the speech of children into text. The KidSense application does not require network connectivity which ensures privacy laws are not breached. This technology can be a very powerful learning tool especially for young children who do not have literacy skills, yet.
This video is intended as an extension activity. Viewing the video is not required.
  • Nuance Dragon
    • Dragon is a speech to text tool which is able to transcribe up to 160 words per minute. This tool can be very helpful for students who have output limitations. Dragon can also be used to enhance spelling ability, increase word recognition, and be used to control a device. The AI aspect of Dragon allows it to learn speech patterns and improve over time.
This video is intended as an extension activity. Viewing the video is not required.

Marking and Feedback

Effectively marking written responses can be a time consuming exercise. The time it takes to effectively mark written responses and provide meaningful feedback can actually be a deterrent to teachers assigning too much of this kind of task. Computers have been able to mark multiple choice questions for a very long time, but this type of assessment is not representative of the understanding that a student has/has not about the content. Written responses in the form of short answer questions or essays are much better at revealing a students level of understanding; however, these types of responses are very labor intensive. In addition to being labor intensive, it can be difficult to provide students with meaningful feedback in a timely manner.

  • Cognii Virtual Learning Assistant
    • Cognii’s Virtual learning assistant is capable of providing individualized learning content but the shining part of this system is its ability to mark open ended questions (short answer). The system is able to not only read the answers for correctness, but also offer feedback for areas to improve.
This video is intended as an extension activity. Viewing the video is not required.

In this context it is worth re-mentioning the work being done in China to mark written responses with AI. China has been working on their AI system for the last ten years which is capable of grading written responses at a 92% agreement with human markers.Though there is no company attached to the project, the technology is being developed.

Tutoring and Learning Support

AI enabled tutoring systems, intelligent tutoring systems (ITS), have the capacity to revolutionize learning. The basic concept is for ITS to monitor how a student solves problems and present additional learning or instruction where needed. Though ITS currently has its limitations, research from Carnegie Mellon University (CMU) is changing how these systems are trained and putting that training into the hands of teachers, not developers. The research from CMU has focused on teaching AI how to teach rather than how to solve. Current AI tutoring systems are very limited in scope and require approx 200hrs of development time to provide one hour of instruction. The new training models from CMU hope to reduced that to a one to one ratio, one hour training for one hour instruction, and put the training of AI in the hands of educators (Byron, 2020). “If this can be accomplished, ITSs will take a significant leap toward being implemented throughout the education sector, online and in classrooms” (Robot, 2020). Current generation ITS systems include Querium, Thinkster Math, and Quizlet.

  • Querium
    • Querium uses AI to deliver customizable STEM tutoring lessons to high school and college students. By analyzing answers and length of time for STEM it took to complete tutoring sessions, Querium’s AI gives teachers insights into a student’s learning habits and designates areas in which the student could improve.
This video is intended as an extension activity. Viewing the video is not required.
  • Quizlet Learn
    • Quizlet Learn, a smart study resource that provides adaptive plans and helps take the guessing out of what to study. The platform uses machine learning and data from millions of study sessions to show students the most relevant study material.
This video is intended as an extension activity. Viewing the video is not required.

Other ITS supported by AI: Thinkster Math

Instructional Content Enhancement and Course Development

AI can be used to analyze student performance, relative to course content, and help to determine where consistent knowledge gaps exist. This information can be used to improve instructional programs.

  • Volley
    • Education technology doesn’t have to be implemented in the classroom. Volley’s AI-based “Knowledge Engine” constantly synthesizes course and quiz results, as well as briefings to find knowledge gaps among employees in enterprises. Companies all over the world are able to quickly and efficiently solve potentially harmful knowledge gaps (lack of general company knowledge, compliance methods or even technical skills) with Volley’s AI.
This video is intended as an extension activity. Viewing the video is not required.

Administration

Most learning systems provide analytics data. AI systems are no different but usually more accurate in their interpretations of data, but that is not what this section is about. A major pain point in education exists around early detection of at risk students. One would think that determining who is at risk and who is not would not be that hard, just look at attendance and marks! Unfortunately, the reality is far more complex. AI models are being developed, alongside data mining of LMS systems, to attempt to determine what an at risk data cluster might look like. The metrics for this work can include things like attendance, grades, family structures, locations, SES, sentiment analysis, learning patterns… As you can see the problem is complex. Though this technology has not been perfected work in this area is underway. If you are interested in learning more about this topic here is a link to a study conducted in Spain which was completed in 2021.

Let’s do a check-in

Have you experienced any changes in education as a result of AI?

How can we make this site better?

Page Resources