How AI impacts today’s classrooms and the exciting path forward
Artificial Intelligence (AI) has a pivotal role in many K-12 educational systems, providing benefits for both students and teachers. To best utilize AI’s potential, it is key for governments to implement policies conducive to AI’s adoption within classrooms. We discuss the benefits and limitations AI provides to education as well as the steps needed to responsibly use AI in education in the future.
AI has contributed massively to making the American educational landscape much stronger and more stable. Through the use of AI in schools, learning has become much more accessible to all groups. With the COVID-19 pandemic devastating the American economy and forcing many schools to move to remote learning, teachers needed to utilize technology, and school districts explored the use of expanding their technological capabilities post-pandemic. This increased use of technology expanded the usage of AI within classrooms.
AI Techniques used in Current Educational Processes
AI can offer many benefits for both students and teachers, making it an important tool in modern education. Thus, AI techniques are being used extensively within classrooms today . For students, AI can mean more personalized learning. Personalized learning is tailored towards a specific student’s needs, and it would allow students to learn the exact content most relevant for them. AI systems can help scale and improve personalized learning. For example, the popular studying tool Quizlet uses AI to help users study more efficiently by creating personalized study paths that best address a student’s weak points, thus helping a student learn content faster . Additionally, similar technology is being used by Duolingo to personalize language learning . Duolingo’s AI system helps users know what they need to practice more and when to practice for optimal results, and it was shown to be highly effective as there was increased user engagement as well as user return increasing by 12%.
Additionally, natural language processing (NLP) techniques have been effective for both students and teachers. Grammarly is known to help improve students’ writing through the use of NLP powered technology. Their systems can analyze text and provide valuable feedback to students, making the quality of their work higher when submitted to teachers. Similarly, the Quizlet system uses NLP to conduct smart grading by checking if a student’s answer matches the correct concept rather than checking if the answers are exactly the same. As a high school student, I have found NLP powered technologies, such as Grammarly and Quizlet, to be quite helpful when studying for tests due to its immediate response system. The technologies are trained to specifically provide aid in struggling sections, which allows me to know which areas I am weak in. For teachers, NLP can help reduce workload which is key since teacher workload has increased tremendously during COVID-19, resulting in teachers risking burnout . Administrative tasks, such as grading assignments or filling out paper, have become burdensome for teachers as they take up a significant amount of time and are largely repetitive. Such tasks can be often automated through the use of NLP systems that can process large amounts of text and give teachers summarizations or complete the task itself .
Similarly, computer vision (CV) techniques, such as object detection or image classification, also have important applications within education. Within schools, safety is an important concern. Beyond directly saving lives, a safer school environment is more conducive to learning. For example, one school district in the United States implemented an AI-based camera system to detect guns . This system can help prevent school shootings, and similar systems can be used to automatically detect unauthorized individuals. As a student in a school that has implemented this system, I feel much safer knowing that the chance of such a tragic event occurring is low, and it makes sense for these types of systems to be implemented all throughout the US.
CV can also greatly benefit teachers. Along with NLP, CV can be used to aid in test grading. For multiple choice tests, CV can automate the grading process by detecting where students have bubbled in their answers . Such technologies have already been implemented within many schools. CV has been a huge relief since it makes test grading much more efficient, and grades are updated much quicker. Additionally, through the use of classification, hand-written characters can be identified, and the resulting text can then be graded. NLP systems can also be integrated in order to parse the text and grade it. CV can also automate attendance processes for teachers. Manually taking attendance takes up important class time while also bringing room for error if teachers incorrectly mark a student as present or absent. Through the use of facial recognition, CV systems can automatically detect which students are present/absent . Additionally, in the case of remote learning, exam security and proctoring was a large concern [9,10]. Through the use of CV techniques like eye gaze tracking and object detection, automated proctoring systems have been made to make remotely-taken exams more secure and ease the burden on teachers who don’t have to manually watch students take their exams online.
AI is also being used to streamline the education process through the use of AI-based tutors and chatbots. Due to teachers being unavailable after school, certain districts have started implementing chatbots that can help improve skills in school-related subjects. In fact, Maryville University implemented a chatbot to answer new undergraduates’ questions regarding the school. Another online tutoring organization called Capacity implemented an AI chatbot that answers students’ questions about specific subjects. Remarkably, it answered 40 student questions at a response rate of 2.7 seconds . The use of an AI chatbot is beneficial in the education space because it saves time for the student and doesn’t require the teacher to work after-school hours to answer questions. Since AI chatbots can be accessed by any device at any time compared to asking teachers and waiting for their response, overall educational output is improved. Despite my high school not implementing a chatbot system, I have had a positive experience with it on tutoring platforms due to its immediate accessibility. Chatbots have massively progressed, allowing me to ask hyper-specific questions such as how to solve certain math equations.
In elementary school, teachers are currently using Presentation Translator, a type of AI tool created by Powerpoint, in which students receive real-time subtitles for everything the teacher says in class, whether it be lectures or just simple instructions for homework . This also increases educational accessibility because English language learners can easily see subtitles in real-time, thus increasing learning and educational output. Through the use of AI, teachers and corporations are both contributing to a better educational space in which students can receive more from their learning.
However, the implementation of AI in education also carries many concerns that need to be addressed for greater adoption. In a world with rapidly advancing technology, data privacy is a problem that is becoming increasingly harder to solve, especially in education. Students use various different technologies in an online or hybrid learning setting that leads to an increase in data collection/usage. While this data is important for the functionality of AI-based applications, it can also result in an invasion of the student’s privacy. In fact, one-third of undergraduate students expressed concern about their privacy when AI technology was used in their classrooms . AI models are trained using user-collected data to create trends and improve the accuracy of the model. Since education technologies are created by large corporations, they strive to collect more data to improve their product. Eventually, these data collection methods seem ordinary, analogous to security cameras, information forms, or usage tracking. Many technologies implemented in schools are pertinent to learning the content of the course, often leaving students no choice to opt-out and remain private. Invasion of privacy can be a major concern for many who wish to remain private and not share much personal data.
Another concern of AI is the expanded inequity between wealthy and underprivileged schools. Due to greater resources, funding, and accessibility, wealthier schools are able to expand AI-based learning faster than underprivileged schools, thus allowing their students to be more technologically proficient. However, underprivileged students may lack technological proficiency, creating a disadvantage for these students if AI is implemented into the school system.
While AI often performs better than the human brain, it lacks self-correction, which is important to improve accuracy. Educational technology could aid towards accurate student placement in a certain subject. These technologies work off the limited data collected while the student uses the application, which could lead to improper outcomes. Generalized algorithms found in AI fail to correctly use small nuances in data, such that they could either inflate the effect of the nuance or completely overlook it. For example, a student could lack vocabulary in a specific topic, and this data could be used to determine the reading level of the student . In a situation where a teacher would be able to identify a student’s strengths and weaknesses effectively, AI lacks the same quality of analysis. While AI may be accurate most of the time, there are many situations in which a human teacher is needed to take a final call.
Currently, minority students face inequality at school, leading to AI models having racially biased data. For example, they are often taught content that they cannot relate to, suspended more often, and placed on decelerated academic paths in contrast to their white counterparts . Language-teaching technologies help build content using standard English, which may not address dialects, sociolects, and slang. AI-based models use this racially biased data to train personalized learning and placement algorithms, thus putting minorities further at a disadvantage. Additionally, underfunded rural school districts may not be able to afford the complex AI security systems or even the teacher-aiding AI systems due to the high price. Since it is a relatively new technology, the price is high which results in its inaccessibility. As per chatbots, many school districts, including my own, don’t have access to this technology, and it can be quite problematic to place all resources on a website chatbot due to some families not having access to the Internet or a device to access it.
Although AI has proved to be beneficial to the educational landscape, it is still in its early stages and much more should be done by the US government in their AI policy-making. The US government should implement policies in which they provide funding to school districts to implement an AI chatbot that comprehends multiple languages. This is important since foreign families move to the US for better educational opportunities, but language barriers can often prevent them from successfully transitioning to the education system. Thus, adding an AI chatbot to school district websites would allow families who aren’t fluent in English to receive information at their own pace and sign their children up for schools or even allow the students themselves to ask about their subjects to the chatbot when the teacher is not active.
The US government also needs to establish data-protection policies that would make the school district’s data collection fully public to all parties. By adopting laws that clearly distinguish between which data can be publicly or privately collected, the government would allow for greater trust to occur which is key for further adoption of AI-technologies in the educational sphere.
As AI becomes more prevalent within the educational realm, it is important to consider both the benefits and concerns of AI usage. Responsible use of AI will allow for maximum benefits for both students and teachers. In the future, AI researchers should consider addressing some of the concerns that exist within today’s models. A move towards privacy-preserving models and less intrusive AI would allow for greater adoption of such systems within classrooms across the world. At the same time, it is important to increase AI literacy among educational administration to allow for greater usage of AI. Additionally, extended applications of current AI technologies could allow for AI to have a greater impact on education in the future, with both NLP and CV having great potential in the educational sphere.
 43 Examples of Artificial Intelligence in Education (2021), University of San Diego
 Hinkle, R, Quizlet has new AI to Help You Study More Efficiently (2020), ELearning Inside
 Peranandum, C, AI Helps Duolingo Personalize Language Learning (2018), Wired
 Radacu, C, Teachers’ Burnout Risk During the COVID-19 Pandemic: Relationships With Socio-Contextual Stress — A Latent Profile Analysis (2022), Frontiers
 Pietro, M, Text Summarization with NLP: Textrank vs Seq2Seq vs BART (2022), Towards Data Science
 Faulkenberry, N. Hobbs school district using AI cameras to detect guns (2022), KCBD
 H. E. Ascencio, C. F. Peña, K. R. Vásquez, M. Cardona and S. Gutiérrez, Automatic Multiple Choice Test Grader using Computer Vision (2021), IEEE Mexican Humanitarian Technology Conference (MHTC)
 Bhavana, D., Kumar, K.K., Kaushik, N. et al, Computer vision based classroom attendance management system-with speech output using LBPH algorithm (2020, Int J Speech Technol
 Prigge, R, Overcoming Remote Proctoring Security Challenges amid COVID-19 (2020), Educause
 Mussachio, M, Secure Online Proctoring For Education: The Threat Of Remote Access (2022), PSI
 Schmidt, J, How Capacity is Powering High Education Institutions: In-Person, Online, and Hybrid (2020), Capacity
 Presentation Translator for Powerpoint (2017), Microsoft
 D. Christopher Brooks, ECAR Study of Undergraduate Students and Information Technology (2016), ECAR
 Torres, J, How to identify, address bias in educational technology (2021), SmartBrief
 Hebbar, N et al, AI in Education Toolkit for Racial Equity (2020), Edtech Equity
Artificial Intelligence is Transforming Modern Education Republished from Source https://towardsdatascience.com/artificial-intelligence-is-transforming-modern-education-d95bf8d19acd?source=rss----7f60cf5620c9---4 via https://towardsdatascience.com/feed