Machine intelligence is changing the future of learning. Here’s what you need to know about how trends like machine learning and natural language processing will shift education.
No doubt, there are a lot of overblown headlines about AI. For example: How Emotion AI Is Making Robots Smarter and Forget The Future, AI Will Take Us Back To The Past.
But when it comes to education, the future is already here. Indeed, many learning tools already use AI.
For instance, many universities now use learning management systems to help guide learners. These tools use AI to categorize students into distinct learning “buckets” so that students can get access to targeted content that is designed according to their abilities and requirements.
Or take the many education nonprofits and companies like Quill using natural language processing to help students learn to write. In short, AI is already changing education. This article attempts to map out the future of the field.
How is AI Helping Education?
One example of the potential power of AI is personalization. For many students, personalization around previous knowledge or interests can have a large impact since students come from various backgrounds and with different interests.
One example is Carnegie Learning’s live facilitation tool, LiveLab. The program won the title for the “Best Use of Artificial Intelligence in Education” in this year’s EdTech Breakthrough Awards.
Designed for grades 6 to 12, the tool offers a personalized learning experience to help students learn at their own pace by taking advantage of big data. It doesn’t only highlight mistakes but it also pinpoints the causes of these mistakes. It also uses big data — and the analysis of that data — to give teachers the ability to see what students are up to in real-time.
Or consider measuring and understanding student emotions. Data on how students feel can be helpful in reducing behavioral disengagement. The tracking of wheel spinning is a good example. Researchers Joseph Beck and Yue Gong define wheel-spinning as “students who do not succeed in mastering a skill in a timely manner.”
Using big data, new intelligent tutoring systems can detect wheel spinning and distinguish between practice which results in “productive persistence” and practice which leads merely to “wheel-spinning.” Understanding how to distinguish between these two kinds of practice within math tools like ASSISTments have helped reduce the number of students who opt out because they have not overcome the process of “wheel-spinning.”
Face-to-face programs have also benefited by prediction tools.
Downsides of Artificial Intelligence in Education
It’s easy to forget that humans are deeply social. We want engagement. The worse punishment besides death for our species is being alone. Plus, some experts argue that it is easier to explain and understand concepts when you’re face to face. There’s also higher motivation through social networks and ties, which often lacks when it comes to AI based learning.
Artificial Intelligence in Education: Beyond The Hype
Natural Language Processing
This technology allows machines to analyze and process large amounts of data related to natural language.
For instance, Duolingo is an app that helps people learn a foreign language and currently boasts more than 300 million users.
The app makes heavy use of AI and claims that 34 hours of Duolingo learning is equivalent to a typical university-level semester. The brand has conducted deep research in the field under the Second Language Acquisition (SLA) model.
Duolingo uses natural language processing (NLP) and AI in a number of ways: to identify the right sentences, to predict learners’ word strength, and to recommend immersion practice documents based on the individual’s progress.
In this way, Duolingo uses AI to create a personalized learning experience. Their website says: “The difficulty of the words, the grammar, and the way we present it to you in the test, all play a role to pick the exact configuration so that in less than five minutes we have a really good sense of where you’re going to start the course.”
Natural language processing can also be used to give feedback. “For instance, NLP can help identify the presence or absence of important discourse elements like claims, arguments, and evidence,” argues professor Scott Crossley. “In addition, NLP can provide feedback to learners about the organization of an essay. These NLP solutions can be combined into automatic writing evaluation (AWE) systems that can provide low level feedback (e.g., tips about vocabulary) or higher level feedback (e.g., advice about the cohesion of discourse.)”
Example of AI-Driven Language Feedback
AI is already changing education. This article attempts to map out the future of the field.
Predictive Tools in Education
But the retention rate in 2015 was as low as 12.6 percent. That’s jaw dropping. The situation hasn’t improved over the past decade, and the dropout rate over the last five years is 96 percent, meaning only 4 out of 100 students complete a course after enrollment.
To address this problem MOOC providers are using AI to understand why students drop out. They study factors like how frequently students check the platform or how quickly they submit assignments. According to an MIT report, it can be difficult to tell who has dropped out until a course ends.
To help reduce the rate, MIT is working on a prediction model that can help institutions predict which students may drop out.
According to some reports, the digital format of the course alone cannot be blamed for a high dropout rate — student competencies also matter. A report found that there are higher chances of a student completing a course if they enroll with a friend. Subsequently, some platforms have begun to offer special discounts on multiple sign-ups.
In an effort to remedy this problem, some platforms have now started to use AI and big data prediction tools to send reminders to students. According to some reports, engagements grow by 30 percent when students are nudged.
Face-to-face programs have also benefited by prediction tools. Berkeley Professor Zach Pardos and his team combine natural language processing and machine learning to tackle problems community college students face as they begin enrollment at UC Berkeley. This ‘data-assistive articulation’ helps alleviate the stress students and administrators face as they attempt to determine which courses should count for credit at the new institution.
The predictive model analyzes the course descriptions, and the order in which courses are taken to give recommendations. It uses this data to provide suggestions for course credit to administrators and students alike.
AI has been a part of learning for a while and many of us have even used the technology without realizing it.
Assessment and Artificial Intelligence
AI allows for assessments to take place in real-time. Experts are working on developing technology that makes it possible for teachers to offer individual student assessments in less time.
Developed at the Harvard Graduate School of Education, EcoMUVE productively uses AI to generate an immersive environment that helps students better understand causal relationships within authentic ecological systems. Designers of the ecosystem use game generated data to provide real time feedback to teachers about students’ attitudes towards science.
In the end, artificial intelligence can help the educational sector develop by making learning easier and more effective.