June 10, 2021
AI in education: 5 ways to reshape the learning experience
Some of you probably remember Professor Keating, the charismatic teacher of the 1989 movie Dead Poets Society who engaged his students with an unconventional didactic approach to promoting their uniqueness.
“Now we all have a great need for acceptance, but you must trust that your beliefs are unique, your own.”
Despite the beauty of its message, certainly shared by many educators, such a perspective seems to clash with the old-style and overly standardized teaching methods we were used to facing at school. However, the legendary professor would have found an unexpected ally some years later. And that ally is artificial intelligence.
If it’s true, as Benjamin Franklin said, that "an investment in knowledge pays the best interest," we should not be surprised by the growing attention, in political and economic terms, given to the synergies between modern technologies and education. Fortunately, I would say, as the teaching environment traditionally suffered from various systemic issues, among which we may include:
In the field of education, the deployment of technologies such as artificial intelligence and data analytics has been carried out with these flaws in mind and with the ultimate goal of correcting them.
Indeed, AI in education today is not just about interactive whiteboards and tablets. Rather, it’s a vibrant environment in which teachers, didactic experts, AI developers, and machine learning consultants work side by side to create innovative solutions and improve the learning experience for both professors and students.
The liveliness of this collaboration is reflected in the growing investments that the sector is enjoying now and will likely benefit from soon. Based on MarketsandMarkets' 2018 AI in Education Market report, the global market is expected to reach $3.68 billion by 2023, growing at a CAGR of 47% during the forecast period.
Such promising estimates are shared by P&S Intelligence. According to their 2020 AI in Education Market Research Report, the global market was valued at $1.1 billion in 2019 and is projected to reach $25.7 billion in 2030 with a CAGR of 32.9%.
The aforementioned study also reports how North America holds the largest market share thanks to the effective blend of massive investments in EdTech and AI, the presence of major market players in the region, a highly developed educational infrastructure, and the growing dependence on tech devices to access teaching programs.
On the other hand, the Asia-Pacific seems to be the fastest-growing market, specifically driven by the massive development of artificial intelligence in China. Previously, the Asian dragon lagged in terms of AI-based adaptive learning technology, as highlighted by Deloitte's 2019 Global Development of AI-based Education paper.
This could be due to the relatively small amount of data collected in the previous years, as opposed to the US and Europe that have been implementing such tools for more than a decade and accumulated information from over 100 million users across different ages. However, the trend will likely change and China is expected to catch up soon, thanks to its large population and skyrocketing development.
Finally, we should not forget about the European Union, which started exploring the use of artificial intelligence and data analytics in education to predict expertise shortages and cope with skill gaps and trends. As stated in its 2018 Digital Education Action Plan (2021-2027), the European Commission considers these technologies as a tool to achieve high-quality, inclusive and accessible education in the Old Continent.
Regarding this matter, another significant feedback by the European Commission comes from the 2018 Impact of Artificial Intelligence on Learning, Teaching, and Education report. This document, drawn up by the Commission's Joint Research Center, identifies the possibility of adapting teaching methods to students' varying needs as the major impact of AI on didactic activities. Therefore, let's start a brief overview of AI use cases in education right from this fundamental point: a flexible and personalized approach to learning.
The latest developments in machine learning (ML), one of the most promising branches of artificial intelligence, have unlocked exciting possibilities in the field of education. Specifically, ML algorithms can process huge data sets, spot specific patterns and relations among all this information, apply mathematical models to represent these connections, and leverage such models to provide suggestions or forecasts. The more data they process, the more these algorithms “learn” and refine their analytical and predictive skills.
The natural result of these capabilities, once applied in the educational environment, is an increasing personalization and flexibility of the learning experience. ML-based platforms can collect student data such as the exercise completion time, test results, and overall performance. Subsequently, this information will be analyzed to understand each student's attitudes and needs, design personalized training pathways, and adapt them in real time by keeping track of the learners’ progress.
For example, if an ML-powered system detects that a student is having difficulty completing a certain task (including tests and exercises), it may advise the teacher to spend more time with that student on this topic. Or, in the case of individual online learning, the platform could offer additional video materials to clarify what the learner struggles to understand. This innovative approach is known as intelligent adaptive teaching and learning.
Nowadays, tools based on the adaptive learning approach are quite common. Carnegie Learning's AI-powered platform, for example, provides teaching and testing to students from kindergarten to college, setting the challenge level based on their current knowledge and skills.
After looking at adaptive teaching platforms and personalized learning, it might be worthwhile to expand this topic further by covering the increasingly popular AI-powered educational tutors and chatbots.
These tools can be integrated into learning platforms to offer more realistic interaction between students and machines. Their ability to communicate with humans is based on speech recognition and natural language processing. Such technologies are typically powered by deep learning (DL), the latest and most powerful variant of machine learning. Deep learning imitates the mechanisms of the human brain to process information via deep neural networks, interconnected layers of artificial neurons that transmit data to each other.
The combination of ML-driven adaptive learning and natural language processing make virtual assistants extremely flexible and, consequently, a valuable ally for learners. Once again, the watchword is personalization: AI-powered tutors and chatbots can design the most efficient teaching style for each student and provide materials and exercises tailored to their needs.
Another strength lies in being operational 24/7 and ensuring a smooth way for students to experiment and learn remotely in a judgment-free environment. Making mistakes while interacting with a bot is surely less embarrassing than doing so in front of your teacher and the whole class.
An example of an AI-based learning platform integrating a smart bot to interact with students is Duolingo’s language-learning application. Its famous tutor/mascot is a green owl that can be as peremptory as Don Vito Corleone from The Godfather when motivating users to making daily exercises.
Duolingo's personalized learning approach includes an initial positioning test to scan the user's skills and a progressive readjustment of the proposed activities' difficulty based on previous performance and results achieved. The ultimate goal is to maintain an optimal level of challenge and keep learners engaged, also thanks to a points-based reward system.
As briefly mentioned above, ML algorithms can examine huge datasets to make predictions about the likelihood of future events. This aspect of artificial intelligence has proved invaluable for data-driven decision-making, even in the field of education.
Think of an ML-based system capable of processing previously collected data on students' academic performance, attitudes, and social conditions and categorizing them into different archetypes based on these characteristics. Subsequently, the algorithms could compare and identify relationships between these categories of learners and their typical schooling outcomes.
The most interesting implication of this process is that the same algorithms will be able, after adequate training on past cases, to detect the same recurring patterns among the newly examined students and predict their future results with a reasonable degree of certainty.
This ability can be leveraged by professors to choose the best teaching methods and programs based on their students' skills and needs. But it's also an invaluable tool for educational institutions, which can leverage predictive analytics software to detect students at risk of dropping out of school or not graduating on time (for example, taking into account poor grades, absenteeism, or tardiness) and support them before it's too late.
A similar solution has been developed by McGraw-Hill Education to integrate it into their AI-powered assessment and learning system called ALEKS. This platform helps teachers identify students who are struggling or at risk of failing and tailor customized learning paths.
Despite the contribution of artificial intelligence in offering an increasingly personalized and adaptive education, the problem of implementing this approach in real life still remains. The main obstacle seems to be the limited time that can be devoted to each student.
According to McKinsey's 2020 How Artificial Intelligence Will Impact K-12 Teachers report, the professionals surveyed work an average of 50 hours a week but spend less than half of this time directly interacting with their students. 69% of them identify the lack of time and flexible schedules as the primary barriers to personalized learning.
However, artificial intelligence could help out in this regard as well. As pointed out by McKinsey, AI and other automation technologies may allow teachers to reallocate 20-30% of their time toward activities focused on actual teaching simply by automating administrative, preparation, and evaluation tasks.
In simple terms, artificial intelligence can help teachers perform the most tedious and time-consuming tasks, such as test evaluation with proper grading software, saving time for interacting with students. Such tools are already perfectly capable of correcting multiple-choice and true-false exercises, but thanks to the developments in natural language processing, they will be more and more efficient in checking short written answers and essays as well.
The Zhejiang International Studies University in east China, for example, has implemented an AI-based tool developed by Alibaba to correct Chinese language tests. This smart checker uses optical character recognition to convert student papers into digital text and then takes care of detecting any errors, including punctuation, redundancy, missing characters, and so on.
Let's close our recap of AI's applications in the education industry with its role in streamlining administrative tasks. After all, education is not just a matter of teachers and students. Many other professionals represent the essential cogs of schools’ and universities’ bureaucratic engines, as they deal with huge amounts of paperwork on a daily basis.
In this regard, AI-driven systems can automate many activities, including the processing of student applications, enrollment, facility management, HR procedures, recruiting, and many more. Indeed, powering administrative tasks and customer service automation with AI may be the decisive factor in containing costs, enhancing management efficiency, and improving responsiveness to students' needs.
That's why the Department for Education of the United Kingdom decided to adopt a system that can autonomously process digital correspondence and, therefore, speed up its follow-up rate to incoming emails. This solution was developed in collaboration with Capgemini and involves robotic process automation technology enhanced with AI.
The growing adoption of AI in education could hold the key to addressing the major flaws that have historically plagued this industry. In fact, artificial intelligence can help to offer a more personalized learning experience, also through the use of dedicated platforms and virtual assistants. Additionally, it can be leveraged to combat school drop-outs and free teachers and administrators from the burden of countless mundane tasks.
On the other hand, AI should be implemented wisely, taking into account its natural limitations and the fact that machines don’t completely replace human action and judgment but, rather, complement it. This assumption is based on some well-known issues:
However, such challenges should not discourage education professionals and institutions from adopting a powerful tool like AI, which is paving the way for numerous new opportunities to be grabbed.
As Professor Keating would say, "Carpe diem. Seize the day, boys".
The role of AI in sports is growing rapidly. Find out how this technology can be leveraged and what challenges its implementation will entail.
One of the main barriers to AI adoption is identifying the right AI use cases. Find out the best practices to make it less challenging.
Learn how automotive industry leverages AI to tap into autonomous driving, increase safety, and enable inter-connectivity.
Find out how machine learning in the stock market is reshaping the way of trading and investing.
Read our guide to supervised vs unsupervised machine learning to choose the right approach for your intelligent software.
Learn about the most relevant RPA use cases by industry and some limitations and requirements to take into account.