A Review of AI Tools: Definitions, Functions, and Applications for K-12 Education

Machine learning

A popular type of AI is machine learning (ML), a technical approach that uses statistical algorithms to build (or learn) a prediction model by processing large amounts of data related to the phenomena of interest [11]. For example, instead of humans programming a set of rules into a system, an AI system of ML can discover patterns among the variables or features on the one hand, and the output variables of interest on the other hand. Two of the most promising applications of ML in education are automated systems that can score student essays, and early warning detection systems that can identify students who are struggling academically and at risk of dropping out [11].

The breakthrough of ML has also given rise to initiatives for teaching ML related principles such as modeling and algorithmic biases in K-12 education [24]. For example, online programs, such as PictoBloc, allow students to take pictures of their index finger as input data to create a model which can recognize their index finger as a way to teach the concept of modeling. Algorithmic biases occur when AI-powered programs make unfair or inaccurate decisions because of biased input data or flawed design. If the students who use the pictures of their index finger in the above example are either white or black, then the model created only using that specific skin color may only recognize a white or black finger and fail to recognize other skin type fingers. Similarly, significant algorithmic biases could occur with facial recognition systems.

Intelligent tutoring systems

Intelligent tutoring systems (ITS) are an early application of AI- powered expert systems in education [11, 25]. ITS can increase student performance by providing individualized learning materials and feedback. ITS can customize the content presented to the student based on his or her level of knowledge in a particular domain, such as in mathematics. DreamBox, an example of ITS also provides the corresponding level of support needed so that students can learn and progress through the content [11]. ITS have the potential to actively engage students in the learning process. For example, students can use tools, such as ChatGPT and other generative AI applications, as personal tutors to complete assignments, study for tests, or learn new topics [25].

Automated essay scoring

Automated essay scoring (AES) is one of the mature applications of AI in education [11]. Teacher’s reading and scoring writing is extremely time-intensive. Thus many teachers are reluctant to assign students extended writing projects. The primary motivation for developing AES applications was to use them for scoring student writing in order to free teachers from the time-consuming task of grading essays [11]. AES can loosely simulate the processing capabilities of the human brain and recognize complex patterns in large multivariate data sets for the purpose of scoring texts. Using AES to score student writing makes such a system valuable to teachers and students because feedback can be given at a much faster rate.

Natural language processing

Natural Language Processing (NLP) is a subfield of AI which has been ubiquitously integrated into education as well as our everyday lives. Applications of NLP include spell checkers, machine translation, recommendation systems, and conversational agents or chatbots [26]. ChatGPT is an example of a useful application of a NLP algorithm allowing users to summarize information, brainstorm ideas, and differentiate lessons and text. NLP can also translate text from one language to another, respond to typed or spoken commands, recognize or authenticate users based on voice, and summarize large volumes of text. Applications of NLP such as chatbots and ChatGPT can also assess the intent or sentiment of a user’s message or speech, answer queries, and generate text, graphics or videos on demand [27].

Improving teaching and learning

The most common function of AI in education is to improve and enhance teaching and learning. The use of AI in education can augment teachers’ capacity by helping them deliver effective instruction [11]. AI enhances teaching and learning by supplementing instruction, personalizing learning, automating grading, and providing adaptive learning platforms that adjust to students’ needs. For example, AI can support teachers with virtual tutors, content creation tools, and data-driven insights to improve teaching approaches. Studies have shown that AI has a positive effect on student learning outcomes. For instance, adaptive learning systems (ALS) and ITS providing personalized teaching and feedback can improve elementary students’ mathematical understanding and achievement [3]. Additionally, AI improves accessibility and creates more effective learning environments, making education more inclusive and responsive to student needs [8]. There are many types of AI tools and products that are currently used to support classroom instruction, including chatbots (e.g., ChatGPT), adaptive learning systems (e.g., Khan Academy), and virtual assistants (e.g., Amazon Alexa) [28].

Personalized or differentiated instruction

The most popular function of AI in education is to provide personalized or differentiated Instruction. Personalized learning, enabled by AI algorithms, tailors the learning process to the unique needs, preferences, and pace of each student [9]. The use of educational AI for personalized learning further enhances the effectiveness of teaching and learning. For example, ITS can model a human tutor in order to improve learning by providing support for the learner, on demand. ALS are personalized learning platforms that adapt to students’ learning progress, the sequence of learning, and the difficulty of tasks. AI-powered chatbots also provide opportunities for personalized learning by engaging students in the learning process, providing immediate feedback, and can be available anytime and anywhere [8, 29].

AI-based solutions have shown promise for supporting teachers in challenging areas of instruction. AI-enabled learning systems have been integrated with existing curriculum to provide more personalized educational settings [30]. Specifically, ALS allow teachers to differentiate instruction at the student level for certain topic areas and skills, and automate scoring of students’ writing assignments to support teachers’ ability to assign more writing projects. Additional and personalized support can also be provided to students based on the early warning function of AI-powered systems to help students stay on track and progress toward learning goals [8]. Finally, scaled-up personalized instructions, driven by innovative AI, enable more students to access high-quality instructions that may not be feasible through traditional in-person methods.

Assessment and evaluation systems (AES)

One desirable function of AI in education is assessing and evaluating students’ learning and performance. AI-supported assessment and evaluation represent an innovative application of AI in education. Automated assessment systems are becoming one of the most prominent and promising applications of ML in K-12 education [31]. These systems play a crucial role in enhancing the personalized learning experience by providing timely assessment and feedback.

For instance, certain assessment systems are capable of instantly scoring students’ writing, exams, and assignments—tasks traditionally performed by teachers and taking a considerable time period to complete [31]. Automated essay scoring (AES) and student writing analysis systems, developed using NLP, assist teachers in scoring texts and writing [11]. AES also extract text features, and through supervised learning with human-scored essays, identify patterns that correlate with different levels of writing quality or score points. Few of these systems are designed to provide holistic scores, while others offer both scores and feedback on individual writing aspects, such as grammar, style, and mechanics [11].

Examples of AES include DreamBox, School AI, and iReady. Use of AES can provide immediate feedback to students and reduce the assessment workload for teachers. Additionally, student engagement is often enhanced by immediate responses and feedback on assignments provided by AES [8, 29].

Early warning or predictive analytics systems

Another important function of AI in education is to predict students’ outcomes and detect failure and student dropouts. Some AI systems have begun to leverage the vast quantity of data to identify at-risk students earlier in their school years [9, 10]. AI can also help those students who struggle in a traditional classroom, by giving them opportunities to engage directly with content in a different approach based on predictive analysis [9]. Using data archived by authorities or higher education institutions from previous cohorts of students, few AI systems employ ML algorithms to identify the most relevant indicators. These algorithms are then applied to data about current students to generate a probability score for each student, typically indicating the likelihood of dropping out of school before graduation [11]. These early warning and predictive analytic systems can help teachers and counselors identify at-risk students early and enable prompt interventions to support struggling students.

In practice, students who struggle with material and assignments may not ask for help, so identification of those at-risk students often happens too late in the semester [10]. ML techniques together with initial assignments and early semester quizzes can predict scores on midterm and final exams [10]. As class sizes increase and teachers’ workloads continue to increase, utilizing AI to help identify struggling students earlier in the semester benefits both teachers and students.

ChatGPT

ChatGPT was developed by OpenAI as a general tool for finding answers, planning itineraries, drafting documents and emails, summarizing and critiquing writing, and creating images, and many other tasks. Sora is a new feature of ChatGPT that creates video from text. There are a lot of discussions regarding the use of ChatGPT in education. Currently, school districts limit the use of ChatGPT on school-owned computers and electronic devices. Special attention should be placed on the ethical and safety measures regarding detection of misleading and hateful content, and algorithmic bias in ChatGPT generated content for young learners [37]. In terms of data privacy, as a ChatGPT user, one can opt out of sharing data with OpenAI for its training model resulting in better data security.

Chat for Schools by Skill Struck

Chat for Schools is an AI platform that assists students and educators in educational settings, offering an environment much like that of ChatGPT. Chat for Schools uses protective walls to keep students’ data safe. While ChatGPT serves a broad range of users for various purposes, from casual conversation to professional writing assistance, Chat for Schools is tailored specifically to the educational environment [38]. Teachers can use Chat for Schools to write lesson plans, gather ideas for class projects, and monitor students’ interactions with ChatGPT [39]. Chat for Schools offers professional development opportunities for teachers and AI lessons for students. Students can use its tutor option to work through problems and assistance with research. All chats are saved for students to refer or revise information discussed with tutors or searched with the system. Teachers can also monitor chats and watch for flagged chats.

Using Chat for Schools, teachers are able to provide their students a safe place to interact with an AI chatbot without jeopardizing their data; receive notification when certain words are detected; have curriculum and professional development available about the risks and opportunities with AI [38]. This platform also gives teachers and students the opportunity to use and interact with AI-powered chatbots with alleviated concerns such as data privacy.

Class Companion

Class Companion is an AI generated feedback tool. It can provide instant and personalized feedback for any written assignments allowing students to improve their writing [40]. Using Class Companion, students have an engaging way to practice writing at their own pace, make mistakes, learn from their mistakes, and improve their writing. Teachers can also track student progress through in-depth reporting on student engagement and mastery of material. Teachers can access built-in resources or generate their own rubrics for feedback.

In terms of ethical concerns and data privacy, Class Companion has a data privacy packet for school districts. Students who are enrolled through their educational institution will have data shared with their specific teacher who can monitor their chats and assignments. Class Companion also follows the COPPA laws for students under the age of 13 years. It is stated that data generated in the system is not sold or provided to third party vendors.

Copilot by Microsoft

Microsoft’s Copilot leverages the power of AI to help teachers improve their productivity as well as creativity [41]. Teachers can use it with a classroom management system to help monitor students’ learning. Teachers can use the Notebook section of Copilot to write a lesson plan and then export the plan. Copilot can help a teacher organize charts or notebooks collected about students. Teachers can also use Copilot when communicating with parents for addressing problems and providing feedback.

Microsoft’s Copilot has a free version [41] for all users. However, some advanced features require specific permissions or subscriptions. When using Copilot, data is anonymized, meaning all personal identifiable information has been removed. Data is not stored indefinitely in the system. Microsoft does not retain prompts or responses from users when using Copilot [42]. Prompts and responses are maintained for a short caching period for runtime purposes. According to Kelli Davis [42], after the browser is closed, the chat topic is reset, or when the session times out, Microsoft discards prompts and responses.

Diffit for teachers

Diffit for Teachers is an AI platform where teachers can differentiate lessons based on existing curriculum or teacher generated content [43]. For example, content materials can be revealed with customized vocabulary, questions, and translations. Teachers can upload a PDF, text, article URL or YouTube video into Diffit which will scaffold it for a specific reading level. This tool can also build background knowledge activities or find lessons generated by verified sources, if required. There is also a growing library of ready-to-be-used activities available for teachers to download. All sources used by Diffit are cited and verified so teachers know the sources of the generated content.

Diffit is FERPA and COPPA compliant and it does not collect student data. However, teachers’ names, email addresses, and IP addresses are collected and provided to Diffit services. Diffit states that the data collected is encrypted and user data is restricted to a limited number of employees at Diffit. Diffit also states that data is not sold or used to advertise third-party products.

Dreambox Learning by Discovery Education

Dreambox Learning is an adaptive platform that can act as a personal tutor for students in mathematics and reading [44]. As a supplemental online program, it provides adaptive instruction for K-12 students. Within the program, the system can adjust students’ learning depending on how they perform. In certain grades, students can have lessons tailored to their proficiency levels.

Potential concerns with using DreamBox Learning include student data privacy and algorithmic bias. The platform collects and stores personal information about students, naturally raising concerns about how securely data is handled and whether it could be accessed by unauthorized parties. Additionally, if the program’s algorithms are trained on biased data (such as only using certain groups of students based on factors like race, gender, or socioeconomic status), the resulting recommendations and learning pathways could lead to some students receiving less effective or appropriate instruction. This could exacerbate existing learning gaps and ultimately affect students’ academic progress.

Eduaide.AI

Eduaide.AI is an AI driven platform that helps teachers create lesson plans and assessments and provides teaching resources [45]. Teachers can revise existing materials for diverse learners to meet students at their proficiency level. The system also provides instant feedback to students so they can revise their writing assignments in any subject area from short answer responses to multi-page essays. Furthermore, curriculum materials in the system can be translated into 15 different languages in order to meet the needs of individual students. Eduaide.AI is mainly for use by teachers and all data is stored in the United States of America.

Khanmigo by Khan Academy

Khanmigo is an AI powered tool by Khan Academy for learning and teaching [46]. Khanmigo offers teachers standards-aligned lesson plans tied to Khan Academy’s content library, Khanmigo-created objectives, rubrics, and exit tickets for teachers. According to Fowler and colleagues [47], exit tickets are short response tasks that teachers administer to students after an activity (such as laboratory experiment) or class period. Teachers and students can also use Khanmigo as a writing coach and personal tutor, as well as prepare for tests. Parents can use Khanmigo to help their children with homework and can access their chat history in Khanmigo for learning difficulties or any inappropriate conversations. Currently, students in schools can only use Khanmigo as a coach if their district has purchased a plan, but teachers can purchase plans to access and use available materials with students.

User information is collected directly when an account is created or a user communicates with Khan Academy or completes surveys. Only users over the age of 18 can purchase a plan using Khanmigo. Parents/guardians, or school districts, need to sign up for anyone under the age of 18 to access Khanmigo. Accounts of young children automatically block features that would allow a child to post or disclose personal information.

Magic Design by Canva

Magic Design by Canva, a design program for creating projects, lessons, and posts, is an AI design generator that creates customized and attractive content in seconds [48]. By using Magic Design templates, students can work collaboratively to create group presentations by adding ideas to be generated. Magic Design gives users the option of letting AI design unique graphics, presentations, and videos from user entered prompts and text. When using Canva, special attention should be paid to the privacy feature of Canva Shield. Canva Shield is the safe environment Canva has built to protect users data while using their app. It also assures users that their data is not used to train their LLM and is kept secure. Canva for Education has a separate privacy policy verifying that no student data is kept by Canva and no advertisements are presented to student accounts. All data from student accounts is in control of the educational institution. Student projects are private unless a teacher chooses to share a link publicly, where it will be shared anonymously.

MagicSchool

MagicSchool, one of the leading AI platforms for schools, helps teachers with lesson plans, feedback, parent communication and a host of administrative and time-saving activities [49]. MagicSchool AI is a suite of tools within MagicSchool designed to leverage generative AI, assisting teachers and students with tasks such as lesson planning, creating assessments, and generating educational content. MagicSchool has recently introduced Raina, an instructional coach chatbot, which offers suggestions when building lesson plans, writing plans, and even gives ideas to improve classroom management [50]. Conversations with Raina can also be saved for review at a later time. As teachers write lessons or look for additional ideas to teach content, Raina can give different ideas and suggestions. Raina’s suggestions for practice are claimed to be research based.

The recent addition of a student version, MagicSchool for Students or MagicStudent, allows students to learn with an AI platform under teacher supervision. For example, students can access a chatbot to ask questions about curricular content and practice taking tests and quizzes based on their interaction with the chatbot over specific material. MagicSchool for Students is independent of MagicSchool for Teachers. With the addition of MagicSchool for Students, comes data privacy concerns with such platforms. The company claims that data collected from students is owned by the schools. Any information that may be used is de-identified and is used for development and improvement to the site in accordance with laws. A Data Privacy Agreement (DPA) is accessible for interested parties using MagicSchool.

PowerBuddy by PowerSchool

PowerSchool’s PowerBuddy is an AI-powered platform for personalized teaching and learning. PowerSchool can also be a grading program where teachers, students, and parents can manage assessment and grades [51]. Teachers can use PowerBuddy to create formative and summative assessments and generate questions to personalize the learning experience for students. Teachers have the ability to create a personalized learning cloud that provides curriculum, information, questions, and assessments based on student interests and needs bases on predictive analytics.

Regarding data privacy, only teachers who have students in class using PowerBuddy are allowed to access students’ data and reports. The parent company of PowerBuddy, PowerSchool, has its own security report which provides its customers transparent information about their cybersecurity, trends in education and cybersecurity, and ways organizations can protect themselves. Data generated in PowerSchool are owned by teachers and schools.

SchoolAI

SchoolAI, advertised as an all-in-one AI platform for the classroom, is designed for K-12 students with built-in monitoring of student’ learning and control of content for teachers [52]. It consists of various tools/functions such as “Build Your Own” for generating specific types of document, “Text Leveler” for adapting texts for specific grade level, and brainstorming tools such as “Story Word Problems” for generating word problems for certain topics. Teachers can use it to generate lesson plans, automate assessment, behavior intervention plans, personalize tutoring, and use data-driven insights to help meet student needs and enhance their learning. The “Spaces” feature offered by SchoolAI is a virtual environment to facilitate learning and collaboration within the platform. Teachers can create their own Spaces for students, focused on specific content. These spaces can be customized for different purposes, such as classrooms, study groups, or projects, allowing students and teachers to interact in a focused and organized manner.

When using SchoolAI, data privacy is a concern, as the platform collects sensitive information about students, teachers, and educational information especially when the information relates to behavior intervention or individualized education plans. Although school districts may have agreements with OpenAI and Microsoft Azure to keep data secure; when teachers use SchoolAI to write lesson plans and behavior plans, the concern will always be where this data is being stored. It is good to know that data generated with SchoolAI is not saved or used by third parties nor used to train models by the system.

Table 1 summarizes the various AI tools discussed and offers practical insights for teachers to consider their use for supporting and enhancing teaching and learning. The tools are categorized by their function, classification, applications, and affordability, enabling teachers to compare and select the ones that best meet their needs. The table provides a go-to reference for teachers who are interested in using AI tools.

Table 1

Educational AI Tools.