Project Requirements
The peer-reviewed project will include five major sections, with relevant sub-sections to organize your work using the CGScholar structure tool.
BUT! Please don’t use these boilerplate headings. Make them specific to your chosen topic, for instance: “Introduction: Addressing the Challenge of Learner Differences”; “The Theory of Differentiated Instruction”; “Lessons from the Research: Differentiated Instruction in Practice”; “Analyzing the Future of Differentiated Instruction in the Era of Artificial Intelligence;” “Conclusions: Challenges and Prospects for Differentiated Instruction.”
Include a publishable title, an Abstract, Keywords, and Work Icon (About this Work => Info => Title/Work Icon/Abstract/Keywords).
Overall Project Wordlength – At least 3500 words (Concentration of words should be on theory/concepts and educational practice)
Part 1: Introduction/Background
Introduce your topic. Why is this topic important? What are the main dimensions of the topic? Where in the research literature and other sources do you need to go to address this topic?
Part 2: Educational Theory/Concepts
What is the educational theory that addresses your topic? Who are the main writers or advocates? Who are their critics, and what do they say?
Your work must be in the form of an exegesis of the relevant scholarly literature that addresses and cites at least 6 scholarly sources (peer-reviewed journal articles or scholarly books).
Media: Include at least 7 media elements, such as images, diagrams, infographics, tables, embedded videos, (either uploaded into CGScholar, or embedded from other sites), web links, PDFs, datasets, or other digital media. Be sure these are well integrated into your work. Explain or discuss each media item in the text of your work. If a video is more than a few minutes long, you should refer to specific points with time codes or the particular aspects of the media object that you want your readers to focus on. Caption each item sourced from the web with a link. You don’t need to include media in the references list – this should be mainly for formal publications such as peer reviewed journal articles and scholarly monographs.
Part 3 – Educational Practice Exegesis
You will present an educational practice example, or an ensemble of practices, as applied in clearly specified learning contexts. This could be a reflection practice in which you have been involved, one you have read about in the scholarly literature, or a new or unfamiliar practice which you would like to explore. While not as detailed as in the Educational Theory section of your work, this section should be supported by scholarly sources. There is not a minimum number of scholarly sources, 6 more scholarly sources in addition to those for section 2 is a reasonable target.
This section should include the following elements:
Articulate the purpose of the practice. What problem were they trying to solve, if any? What were the implementers or researchers hoping to achieve and/or learn from implementing this practice?
Provide detailed context of the educational practice applications – what, who, when, where, etc.
Describe the findings or outcomes of the implementation. What occurred? What were the impacts? What were the conclusions?
Part 4: Analysis/Discussion
Connect the practice to the theory. How does the practice that you have analyzed in this section of your work connect with the theory that you analyzed on the previous section? Does the practice fulfill the promise of the theory? What are its limitations? What are its unrealized potentials? What is your overall interpretation of your selected topic? What do the critics say about the concept and its theory, and what are the possible rebuttals of their arguments? Are its ideals and purposes hard, easy, too easy, or too hard to realize? What does the research say? What would you recommend as a way forward? What needs more thinking in theory and research of practice?
Part 5: References (as a part of and subset of the main References Section at the end of the full work)
Include citations for all media and other curated content throughout the work (below each image and video)
Include a references section of all sources and media used throughout the work, differentiated between your Learning Module-specific content and your literature review sources.
Include a References “element” or section using APA 7th edition with at least 10 scholarly sources and media sources that you have used and referred to in the text.
Be sure to follow APA guidelines, including lowercase article titles, uppercase journal titles first letter of each word), and italicized journal titles and volumes.
Assessment of student progress in the classroom has always been a difficult aspect of teaching. Even after fifteen years of teaching, I still struggle with the balance of assessment: we need assessments in order to determine student mastery, but punishing low-performing students with low test scores is a good way to demoralize and disengage them. More creative assessments that allow students to demonstrate their mastery in more genuine contexts like projects and discussions are certainly better at engaging students, but they’re also much more difficult to get accurate data from: students are typically clever enough to work around knowledge gaps when completing projects, and then their teacher may not even realize those knowledge gaps are there. Consistency can also be a major challenge in assessing students, particularly in my field of English Language Arts, as even the most specific rubric in the world could not fully safeguard graders from all of the external forces that may influence the score we give a particular essay on a particular day.
In a previous course in Scholar, I created a work focused on the utilization of AI in the ELA classroom, which covered a number of possible projects and activities an ELA teacher could use with the help of an AI in order to enrich their students’ learning. In this work, I would like to focus on AI’s utility in assessment. Harry and Sayudin (2023) write that, “automated grading and assessment can increase efficiency, save teachers’ time, and provide more accurate and consistent feedback,” (2023). Certainly, grading student work–particularly long assignments like essays or projects–is time-consuming, but I think the endurance of an AI grader is at least as important as its speed. There have been days when I have looked upon a stack of papers needing to be graded with dread. Did my exhaustion impact the grades students received on those papers? Possibly. An AI grader, immune as it is to headaches, would not have such problems, and, importantly, a teacher whose grading is completed automatically would have more time and energy to dedicate to the learning of his or her students.
Recent advances in Artificial Intelligence have made such integration very feasible. In their study on AI integration in education, Holmes and Tuomi (2022) write:
“Research at Stanford University evaluated an autograder AFA (Automatic Formative Assessment) system that provided feedback on programming tasks completed by 12,000 students in a computer science class. The students agreed with the given feedback about 98 percent of the time, slightly more than their agreement with feedback from the human instructors,” (2022).
AI-driven grading is already technologically possible and generally reliable, and given how quickly AI technology is improving, likely to only get better. The question, then, is not whether an AI can be used to grade student assessments but whether using an AI tool to do so helps student learning.
Harry & Sayudin (2023) write, “AI can provide better data analysis, enabling educators to make data-driven decisions,” (2023). Such data analysis is invaluable to educators. When I moved from teaching in person to teaching online during COVID, I found the amount of data the LMS stored to be staggering–and certainly much more detailed than my physical gradebook–because the digital environment my students were working in stored all of that data automatically. For any given student, I could access half a dozen charts and graphs that broke down just about everything there was to be broken down about their performance in class. If a student’s grade was low, I could use that data to find patterns in their performance to determine where their knowledge gaps were and create interventions to match.
AI, I believe, can take that even further. An AI could not only collect that data but also analyze it, giving a breakdown on student performance in any number of defined categories, allowing an educator to read a summary of a student’s progress over their past several AI-graded assessments. The AI could find and summarize the student’s knowledge gaps, which would allow the teacher to focus their time and energy on interventions and reteaching.
In their article “Artificial intelligence for education: Knowledge and its assessment in AI-enabled learning ecologies,” Cope and Kalantzis (2021) write that these utilizations of AI “offer the potential to transform education in ways that—counterintuitively perhaps—make education more human, not less,” (2021). I do not believe that it is my grading skill that makes me a good teacher, nor do I think students find any educational value in my grading of their assessments. Instead, it is the interactions I have with my students based on the results of their assessments that help drive their learning, and utilizing AI tools to expedite the back-end work of assessing student performance would be a tremendous help in allowing me to focus my efforts on those interactions.
The time commitment required for a human teacher to grade assessments and provide thorough enough feedback on them to be helpful to student development is nontrivial, as well. In my own teaching, I have encountered the conundrum many times in which I want to provide my students detailed, constructive feedback, but by the time I've been able to do so for the hundred and fifty or so students I have at a time, we've moved on to a new lesson in class. Do I then give students their assessments back, tell them to read my feedback and make a note of it for the next time a similar assessment comes up, or do I interrupt the new lesson we've started to backtrack to the previous one for students to practice the feedback I've given them? An AI assessment tool, I believe, could be instrumental in solving this dilemma by grading assessments and providing feedback on student mastery to the teacher very quickly, which would allow students to incorporate the teacher's feedback and practice the suggested strategies while they are still working on the relevant lesson.
Thus, in this work, I will be focusing on exploring the utilization of Artificial Intelligence in student assessment. How can assessments be designed in such a way that creates a well-rounded picture of student mastery for an AI to analyze and summarize (and, ideally, that doesn’t make students hate school in the process)? How can AI assessment be utilized to further drive student learning and improve engagement? What other utility can AI provide in terms of assessment and post-assessment intervention? As a high school English Language Arts teacher, my focus will be on that particular age group and subject area, but I believe these findings will have more universal applications, as well.
Research around the use of Artificial Intelligence–and particularly generative AI–as an assessment tool in education is, by the simple newness of the technology, relatively new and very much ongoing. However, research into the impacts and potential impacts of machine learning on education in general and assessment specifically goes back at least two decades. Given that generative AI is really just the most recent and advanced version of machine learning we have available, pedagogies based around the use of machine learning algorithms can be extrapolated to apply to generative AI in many situations.
Interestingly, the use of generative AI as an assessment tool in education fits very well with a constructivist learning environment. Constructivism emphasizes the importance of putting students in the role of active learners, where they work through experiential learning to develop a well-rounded understanding of topics by integrating new information with prior knowledge (Larochelle et al. 1998). In the constructivist classroom, formative assessment–that is, iterative assessment done during the learning process–is more important than a summative assessment–assessment done after the learning is complete–because constructivism focuses on the iterative nature of education: students learn a concept and that learning becomes a foundation for new learning (Larochelle et al. 1998).
Generative AI fits into this model because it is able to combine near-instant assessment with detailed and thorough feedback. Grubaugh et al. (2023) write:
“This learner-centered approach [constructivism] aligns with intelligent tutoring systems and other AI applications that adaptively respond to students' existing mental models to promote deeper learning. Educators, often unknowingly, employ a range of artificial intelligence tools embedded in everyday software, such as Microsoft Word, as used in this article, information retrieval, and numerous other tasks. Teachers can now harness AI tools like ChatGPT, BARD, Microsoft Bing, and others, deliberately integrating them into constructivist pedagogy to bolster student engagement, metacognition, and conceptual change – all while upholding the humanistic values of education,” (2023).
To me, as an English teacher, the essay is the perfect example of an educational practice that can benefit from such AI usage. In the traditional classroom, the teacher assigns an essay on Monday, collects the completed assignments on Friday, grades them over the weekend, and hands them back to the students the following Monday. From the student perspective, they were taught a lesson on writing, given an assignment using the skills learned in that lesson, completed that assignment mostly or entirely on their own, and then received a grade for it a few days later. The teacher will likely have given them some feedback on their writing–work on your thesis statement, avoid “in conclusion” and other trite transitions–but the students won’t have a chance to implement that feedback into their own learning until the next essay comes around, likely several weeks later, and the process begins again.
This, in my experience, does not accomplish the goal of teaching students written communication skills in any way that could be called satisfactory. There are many complex skills that go into the writing process, especially the essay-writing process, and students are expected to learn and master all of them in a single assignment. This can be broken up into smaller chunks, such as having students submit outlines and rough drafts, but it is still an enormous task for students who are likely overwhelmed by the process. A constructivist would say that the solution is to have students write their essay in parts, iteratively, receiving feedback on each submission until they feel they have mastered that particular skill before moving on to the next and repeating the process until they have developed a full and well-rounded understanding of essay writing.
For a single teacher who may have thirty students in a single class and five of those classes in a single day, such a process is unfeasible. An AI assessment tool, however, could provide just that: rapid assessment of iterative student work and feedback on what each student needs to work on. In their article on using machine learning in student assessment, Cope and Kalantzis (2015) write:
“Indeed, it is conceivable that summative assessments could be abandoned, and even the distinction between formative and summative assessment. Take the practice of big data in education, or the incidental recording of learning actions and interactions. In a situation where data collection has been embedded within the learner’s workspace, it is possible to track back over every contributory learning-action, to trace the microdynamics of the learning process, and analyze the shape and provenance of learning artifacts,” (2015).
If properly utilized, generative AI can not only be used to complete such tasks but also to provide feedback reports to teachers for them to use in targeted interventions. Imagine, then, that the class assigned the essay now has access to an AI assessment tool. Rather than being expected to complete the entire essay process in one submission, the students complete smaller, more targeted writing tasks: create an interesting and contextualizing introduction paragraph, develop a concise thesis statement, integrate conceptual transitions, provide sufficient detail in body paragraphs, and so on. As they complete each, the AI tool provides iterative, personalized, formative feedback on their submissions that the teacher can then use to implement targeted interventions to help the students master the concepts. This allows students to build their understanding of essay-writing piece-by-piece to create a fuller mastery of the skills involved, rather than overloading them with all of the information at once and hoping they remember most of it when the due date comes around.
There are a number of crucial criticisms of generative AI tools to take into account when implementing such tools into the classroom, such as equity issues around race and socioeconomics as well as school and student access to these technological tools. This work, therefor, will also focus on exploring those criticisms and addressing them in their applications.
The majority of criticisms of artificial intelligence focus on the ethics of both its creation and its usage. Who is involved in the development of AI algorithms and what does that mean for the groups of people who are not involved? If AI pulls from otherwise protected sources online–particularly art–what does that mean for the original creators of those sources? If the only consideration of an AI is the “averaging” of all information it pulls from, does that further marginalize already marginalized groups, and does it or can it propagate misinformation? As AI tools continue to be developed and refined, a field of ethics is being developed around it, and these are only a few of the wide-reaching questions that can be found within that field.
For the purpose of this work, I will be focusing on those criticisms that–at least from my perspective, as a teacher–are the most applicable to the use of AI as an assessment tool in education. While many other ethical considerations could and likely do have some impact on this usage, they are largely beyond the scope of this work.
One of the most crucial aspects of student assessment is the question of fairness. Several factors go into the decision of how one student is assessed compared to another, including special accommodations, student development and mastery levels, level of effort, and so on. An essay from a student in an honors course is likely going to be graded more strictly than one from a student in a remedial course, for example, and teachers typically have a sense of the capabilities of the students they have in a class they can pull from when grading. In most situations, student assessment is not a competitive issue but a constructive one, and the assessment that best helps the student learn and grow is typically more useful than the one that is rigidly equal to every student it assesses.
An important criticism of artificial intelligence usage in education is that it does not have the sense of fairness a human teacher would have. Simon Lindgren (2023) writes:
“...there can never be any ‘complete AI fairness’. AI is not some magic, superhuman, system that we can tweak to some final and optimal setting. Rather, the results and societal consequences produced by AI are fundamentally constrained by the convictions and expectations of those that build, manage, deploy and use it. What is seen as fair by some people may not be perceived to be fair by others, and the complex networked connections among people and technology in society mean that what ensures fairness for one group may produce unfairness for others,” (2023).
In many ways, this perspective of fairness applies to the concept of assessment, as well: a grade that is fair to one student may not be fair to another. The difference is the human element. When grading student work, teachers are able to use their professional judgment and experience to determine what is fair for each student. An artificial intelligence assessment tool would likely lack that level of understanding, and so it would need to be implemented in a way that mitigates that lack.
The question of equity in artificial intelligence, then, is at least as important in its educational applications as it is in other fields. AI tools are typically developed by powerful tech companies like Microsoft and OpenAI, and so the “perspectives” of those AI tools often reflect those of people in powerful positions. Critics argue that this creates an intrinsic power imbalance in the use of artificial intelligence.
Ricaurte and Zasso (2023) write that, “ethical frameworks developed by powerful actors are problematic if they do not address power asymmetries and structural and epistemic violence,” (2023). How would an AI assessment tool developed by an ultra-wealthy tech company grade the writing of a student from a low socioeconomic status (SES) household compared to one from a high-SES household? Intrinsic bias can be inherited by an AI from its developers in subtle ways, which can surface as micro-or-macro aggressions against people of different ethnic, wealth, language-use, and other backgrounds (Ricaurte & Zasso 2023). An effective AI assessment tool, then, would need to compensate for such issues in order to be fair, equitable, and helpful to students.
Another important criticism of implementing AI tools in education is the same as many criticisms of other educational technologies: does the implementation of this tool benefit students and teachers, or is it simply being implemented because it is a new technology? Holmes and Porayska-Pomsta (2023) write:
“The shift from research labs to real-world – and usually commercial – applications, accelerated by the COVID-19 pandemic, means that AIED (artificial intelligence education) is fast becoming a tool that sets the tone for how educational policies are enacted worldwide. Accordingly, it is increasingly important for AIED researchers and developers to stand back, to think about how their technologies are imposing on or contributing to the system, to consider what type of world they are helping to create, piece by piece, with “small acts of technology-based automation” (Selwyn et al., 2021, p. 1). If the research or development fits or serves the current system, despite the fact – and knowing that – it is broken, while understanding that AI has the capacity to amplify, reinforce, and perpetuate poor pedagogic practices – is this ethical?” (2023).
Implementing any technological tool to streamline a bad practice does not make that practice suddenly better. A common example of this is the Learning Management Systems (LMS) used by many online education companies: rather than using digital technologies to innovate new and helpful educational practices, many of these LMS’s simply take old pedagogies–like lectures followed by tests–and prop them up in a digital space (Selwyn et al., 2021). This does nothing to help students grow and learn; they are simply experiencing the same didactic pedagogies of the mid-20th century but in a digital classroom instead of a physical one. Thus, in order for an AI assessment tool to be helpful and useful, it needs to do more than complete an existing task marginally better: it must address a need in schools in a way that genuinely uplifts the education of students and the performance of teachers.
Many of these issues are daunting, and while it is beyond the scope of this work to address the code-level development of artificial intelligence tools, the goal of presenting applications for these tools in ways that address and mitigate these criticisms is, I believe, a very reasonable one. In the following section, I will work to synthesize the educational theories supporting the use of AI assessment tools in education with the criticisms of the technology presented here to create such applications.
Implementing an AI assessment tool into a classroom is not something that should be done lightly. Many implementations of new technology into education fail because the goal of uplifting student education is lost in the face of the technology’s convenience or simple newness; digital education spaces that are little more than platforms for hosting lecture videos and multiple-choice tests are a prime example of this. Rehashing old pedagogies with new technologies does nothing for students and is often worse for teachers.
It is also important for teachers utilizing AI tools to understand how those tools work, what their limitations are, and which tools are best for which situations. Machine learning tools are used to recognize, respond to, and create patterns, and so they could be a better match for an assessment that requires students to create and label a map, for example. An artificial intelligence tool, on the other hand, is able to utilize logic to create complex analyses and generate detailed responses. Because of the complexity of these tools, it is important that teachers receive sufficient training to understand best practices around their use, particularly in terms of their limitations; a lack of understanding of such a tool can lead teachers to use them for tasks they are not designed for, which could harm, rather than help, student development.
Artificial intelligence–along with other machine-learning and algorithm-based tools–also comes with its own set of ethical concerns. Although these technologies attempt to achieve a “default” perspective in their output, in truth there is no such thing; the backgrounds and life experiences of the developers of these technologies impact how they interact with people from different backgrounds and life experiences than they have had in ways ranging from the subtle to the blatant. Much like existing algorithms, AI tools can be unfair to minority groups, which raises significant ethical concerns over the equity of these technologies when they are used broadly.
The successful implementation of AI assessment tools in education, as is the case with successful uses of AI in other fields, relies on the human element. Cukurova (2024) writes:
“Humans are very good at many things that today's AI is still pretty poor at, and AI is good at some others. Machines are much ahead of humans on some variables like computing floating point arithmetic, yet way behind on others like cognitive flexibility and long-term planning in unusual situations. This is not to say that humans are more intelligent than machines, or vice versa, we are differently abled,” (2024).
In the realm of education, this can be taken to mean that AI tools are good at collecting data while human educators are good at using that data to educate students. Applying this dynamic to assessment, then, is fairly straightforward: a human teacher could take hours to grade all of his students’ essays and still not collect as much data in as organized a way as an AI grader would in a couple of minutes. However, it is the cognitive flexibility and long-term perspective of the human educator that would best apply that data to directly interact with students. Cukurova (2024) continues:
“Recognition of these differences in abilities provides a strong argument for the value of hybrid intelligence systems which are tightly coupled human–AI systems where both entities interact smoothly and dynamically, leveraging their respective strengths. In educational contexts, hybrid intelligence systems can significantly enhance human competence development by combining human cognitive flexibility, reflective long-term planning and real-world contextual understanding with AI's data processing capabilities and analytics,” (2024).
This model helps to avoid the pitfalls of AI algorithms that are the subject of much of the criticism surrounding AI, such as its inherent bias against minority groups and inability to incorporate contextual information into its processing, because the AI is not directly interacting with the student as part of the assessment. Instead, the human educator applies their contextual understanding of her students to the data collected by the AI assessment tool in order to create interventions for the students that are appropriate for their mastery levels and circumstances.
For example, consider a student in an English class whose first language is not English. Their writing mechanics in English might not be as developed as their native-English-speaking peers' and so an AI assessment tool, lacking that context, may penalize that student harshly for their mechanical errors. However, that student's human teacher would recognize that their language mastery is at a different level, and rather than use the AI assessment tool to provide a grade to the student, they could use it to target specific areas in the student's writing that need improvement to help them develop those skills. That way, the student is still receiving constructive feedback but is also receiving encouragement to improvement rather than punishment for their mistakes.
Similarly, a student in an American History class who is not originally from the United States may respond to prompts in that class from a very different perspective than those of their peers. An AI assessment tool, if used to provide feedback to students directly, may very well mark those perspectives as wrong and deduct points, given the cultural context that the AI tool's developers likely had in creating it. A human teacher, however, could pull targeted, constructive feedback from the AI assessment tool's report and incorporate it into their own feedback to the student in a way that allows for and encourages multicultural perspectives on topics.
However, I believe that AI assessment tools can also be used to provide iterative, specific feedback directly to students for tasks related to correct-or-incorrect skills, so long as that feedback is used constructively and not as actual grades that could penalize the students for errors. Writing mechanics skills are a good example of a skill set that an AI could directly interface with students on in order to develop those skills iteratively. Owan et al. (2023) write:
“AI can provide feedback on grammar, spelling, and syntax by analyzing essays, reports, and other written assignments. By using automated grading systems, teachers can focus more on essential tasks such as lesson planning and supporting students, resulting in significant time savings. This can help students to improve their writing skills and reduce the workload for teachers,” (2023).
This type of direct-to-student assessment is low-stakes because the AI’s feedback is not directly reflected in the students’ grades, but it also fits well into the iterative, formative assessments of a Constructivist classroom by providing immediate and constructive feedback to students as and when they complete and revise assignments. This model incorporates the utility of AI processing power while mitigating many of its pitfalls because it does not need to have contextual information about a student in order to give that student feedback on specific, correct-or-incorrect type tasks.
In order to utilize an AI assessment tool effectively, however, its sources will also need to be controlled. Tzirides et al. (2023) write:
“To validate antecedent knowledge claims, we need to be able to interrogate their sources. To distinguish the thinking of the writer from the social knowledge upon which that thinking is based, we use quotation and citation apparatuses. In school, we call this “critical literacy.” In academic work, the credibility of sources is dependent on a number of variables, including the qualifications of the researcher, the credibility of the publication venue, and the rigors of peer review. However, giving us the impression that the AI is answering rather than its sources, the sources are hidden. This makes the AI seem smart when it is just a remix of sources. Some of the sources, moreover, may be embarrassing to disclose; others have been copied in breach of copyright,” (2023).
Many publically-available generative AI tools, such as ChatGPT, pull from wide swathes of online sources; this provides a widely-sourced “average” of information in its output but also means that it can give information that is wrong simply because there is a lot of wrong information on the Internet (Tzirides et al. 2023). Using this type of “averaged” data extensively in education also runs the risk of building student knowledge that is too generic or unspecific, rather than allowing for the true development of student mastery of subjects (Tzirides et al. 2023).
In many ways, incorporating human educators in the assessment process mitigates these drawbacks considerably: a human teacher would be able to contextualize data provided by an AI assessment tool and create a more genuine learning environment for the students based on that data than an AI would be able to. However, it is also important to ensure that the feedback provided both to students and to teachers based on these assessments is accurate, constructive, and targeted. One possible solution to this issue is to develop the AI assessment tool’s constructive and contextual understanding of the subjects it is assessing by feeding it quality responses to assignments within that subject as part of its prompt engineering. On developing the AI feedback tool for the CGMap learning platform, Tzirides et al. (2023) write:
“What we need for reliable knowledge work and good learning is to feed the machine with the epistemic virtues of using reliable sources and resilient facts, theories, and critical perspectives. We do this in CGMap in two ways. First, the generative AI is presented student texts that have already been vetted by peers for these epistemic virtues. Then, second, we use generative AI to provide supplementary narrative reviews through careful prompt engineering. In a sense, we have told the GPT in general terms what to say in response to the specifics of the student texts,” (2023).
Structuring an AI assessment tool in this way would empower it to avoid many of the informational and factual pitfalls often found in generative AI tools as well as to ensure the feedback and data it provides is targeted and specific. In many ways, such targeted prompt engineering will help to avoid many of the bias issues present within AI tools as well, because the data it is pulling from comes directly from the students and teachers of the school or other learning environment where it is being used. This creates, if not a genuine context for the AI tool, at least a context that is closer and more relevant to the students and teachers it is being used to assist.
Incorporating these tools into schools effectively would be a challenge. The task of acquiring funding for such tools alone is nontrivial, given the relatively low budget most schools have to work with. The school where I had my first teaching job, for example, provided me with a budget of about fifty cents per students per semester, and that was supposed to cover all of the supplies I might need at that time; as an English teacher, that barely covered paper, pencils, and a few treasured poster boards. While addressing funding issues is outside the scope of this paper, those issues still represent significant hurdles to the implementation of AI assessment tools.
Teachers will also need to be trained to use these tools effectively. Without training, teachers are much more likely to use complex digital tools incorrectly or not at all; I have seen firsthand technology initiatives fail because they only provided teachers with the tools and not the training to use them or lessons that incorporate them. Sufficient teacher training is ultimately a funding issue, but it is also an issue of deliberateness: administrators and policymakers must understand that artificial intelligence is a complex tool with many uses and potential pitfalls, and must therefore provide training and resources accordingly.
As a teacher, I genuinely believe that AI assessment tools, if used correctly, could empower teachers to drive their instruction with more targeted and specific data as well as provide them with more time to focus on interacting with their students by completing many of their time-intensive, backend tasks for them. Rather than dehumanizing the learning process, I believe this will further humanize education, as the AI assessment tools would be used to handle the tasks that typically take teacher time away from student interaction and free them up to spend more time on educational activities. It is important, however, to note the many possible pitfalls and drawbacks of AI tools and work to incorporate those tools in a way that mitigates or eliminates them. AI assessment tools, then, must be both structured and incorporated deliberately, with specific goals in mind, and not haphazardly integrated into schools without a plan as many educational technologies have been. Artificial Intelligence tools are only becoming more advanced and more prevalent, and so deliberate methods of incorporating them into education are crucial in order to avoid AI implementations that do more harm than good.
Akgun, Selin & Greenhow, Christine. (2022). Artificial intelligence in education: Addressing ethical challenges in K-12 settings. AI and Ethics, 2. 10.1007/s43681-021-00096-7.
OpenAI. (n.d.). Chatgpt. ChatGPT. https://chatgpt.com/
Cope, B., & Kalantzis, M. (2015). Assessment and pedagogy in the era of machine-mediated learning. In T. Dragonas, K. J. Gergen, S. McNamee, & E. Tseliou (Eds.), Education as social construction: Contributions to theory, research, and practice (pp. 350–374). Worldshare Books.
Cope, B., Kalantzis, M., & Searsmith, D. (2021). Artificial intelligence for education: Knowledge and its assessment in AI-enabled learning ecologies. Educational philosophy and theory, 53(12), 1229-1245.
Cukurova, M. (2024). The interplay of learning, analytics and artificial intelligence in education: A vision for hybrid intelligence. British Journal of Educational Technology.
Grubaugh, S., Levitt, G., & Deever, D. (2023). Harnessing AI to power constructivist learning: An evolution in educational methodologies. EIKI Journal of Effective Teaching Methods, 1(3).
Harry, A., & Sayudin, S. (2023). Role of AI in Education. Interdisciplinary Journal and Humanity (INJURITY), 2(3), 260-268.
Holmes, W., & Porayska-Pomsta, K. (2023). The ethics of artificial intelligence in education. Routledge Taylor.
Holmes, W., & Tuomi, I. (2022). State of the art and practice in AI in education. European Journal of Education, 57(4), 542-570.
Larochelle, M., Bednarz, N., & Garrison, J. W. (Eds.). (1998). Constructivism and education. Cambridge University Press.
Lindgren, S. (2023). Introducing critical studies of artificial intelligence. In Handbook of critical studies of artificial intelligence (pp. 1-19). Edward Elgar Publishing.
Mangino, M. E. (2019). Constructivism. https://manginodesign.mit.edu/comparing-learning-theories/constructivism
Owan, V. J., Abang, K. B., Idika, D. O., Etta, E. O., & Bassey, B. A. (2023). Exploring the potential of artificial intelligence tools in educational measurement and assessment. Eurasia Journal of Mathematics, Science and Technology Education, 19(8), em2307.
Ricaurte, P., & Zasso, M. (2023). AI, ethics and coloniality: A feminist critique. In M. Cebral-Loureda, E. G. Rincón-Flores, & G. Sanchez-Ante (Eds.), What AI can do: Strengths and limitations of artificial intelligence (pp. 53–72). Milton Park.
Selwyn, N., Hillman, T., Bergviken Rensfeldt, A., & Perrotta, C. (2021). Digital technologies and the automation of education: Key questions and concerns. Postdigital Science and Education.
Spencer, J. (2023, October 2). 5 ways artificial intelligence might transform assessment practices. John Spencer. https://spencerauthor.com/ai-assessment/
The Peak Performance Center. (n.d.). Constructivism. The Peak Performance Center. https://thepeakperformancecenter.com/educational-learning/learning/theories/constructivism/
Tzirides, A. O., Zapata, G., Saini, A., Searsmith, D., Cope, B., Kalantzis, M., Castro, V., Kourkoulou, T., Jones, J., Abrantes da Silva, R., Whiting, J., & Kastania, N. P. (2023). Generative AI: Implications and applications for education. arXiv. https://doi.org/10.48550/arXiv.2305.07605
Zywica, J. (2023, May 4). Effective feedback loops help students engage and learn actively. Opportunity Education. https://opportunityeducation.org/white-paper-effective-feedback/