AI in Student Assessment and Feedback

Introduction: A new paradigm in assessment and the redefined role of the teacher

The entry of artificial intelligence (AI) into education is not just the introduction of another technological tool; it is a catalyst that forces us to re-evaluate the very foundations and principles of assessment. Artificial intelligence does not replace the teacher but fundamentally transforms their role. From the position of the primary provider of information and the final arbiter of assessment, the teacher moves into the role of an architect of educational experiences, a mentor, and an ethical guide in an increasingly complex digital world.¹

This chapter deals with the use of AI in assessment and providing feedback, drawing on a humanistic approach also emphasized by international organizations like UNESCO. The goal is for technology to serve to enhance human capabilities, support autonomy and critical thinking, and not to replace or weaken them.³ AI must contribute to fair, inclusive, and equitable education for all.

The current situation is defined by a fundamental tension. On one hand, AI offers enormous potential for efficiency and personalization that we could only dream of in the past. It allows for the automation of routine tasks, providing instant feedback, and creating individualized learning paths for each student.⁵ On the other hand, it brings serious pedagogical and ethical challenges. These include new forms of academic dishonesty, the risk of eroding key cognitive skills, and the threat of systemic discrimination through algorithmic bias.⁷

The main task of the modern educator is thus not just to passively accept new tools, but to actively and consciously navigate this tension. The advent of generative AI, such as ChatGPT, changes the dynamics of the relationship between the teacher, the student, and technology. AI is no longer just a fast calculator for grading tests; it is becoming a partner capable of conducting dialogue, generating content, and analyzing arguments.¹⁰ This shift requires a fundamental change in approach - from assessing finished products to assessing the thinking process and teaching students how to work with this new technological "partner" effectively, critically, and ethically. This chapter serves as a guide for teachers on how to maximize the benefits of AI while consciously and actively minimizing its risks.

The potential of artificial intelligence to transform assessment

The integration of artificial intelligence into assessment processes opens the door to fundamental changes that can increase the efficiency, fairness, and personalization of education. The potential of AI lies in three key areas: automation, which frees up teachers' time; personalization, which adapts feedback to the needs of the individual; and data analysis, which provides valuable information for improving instruction itself.

Automation and efficiency: More time for what matters

One of the most immediate and visible benefits of AI in assessment is the ability to automate routine and time-consuming tasks. Teachers spend a significant portion of their working time correcting tests, checking homework, and grading written assignments.¹² AI can significantly streamline these processes. Tools like Google Forms with integrated AI extensions can automatically grade tests with closed-ended questions, giving the teacher an immediate overview of the class's results.⁵ Applications like Grammarly or LanguageTool can provide students with instant feedback on the grammatical and stylistic correctness of their texts, thus relieving the teacher of correcting basic errors.⁵

The initial benefit of "time-saving" is obvious, but the deeper meaning lies in how this saved time is reinvested. The goal is not to work less, but to work differently and more meaningfully. The time a teacher would have spent grading multiple-choice tests can be strategically invested in activities that AI cannot replace. These include conducting an empathetic conversation with a student struggling with the material, moderating a complex class discussion, providing nuanced feedback on a creative project, or simply giving more individual attention to those who need it most.¹ Automation is thus not an end in itself, but a means to strengthen the irreplaceable human dimension in teaching.

Personalization in practice: Immediate and targeted feedback

The traditional assessment model is often based on a "one-size-fits-all" approach, where all students receive the same tasks and are assessed by the same criteria at the same time. AI allows for a transition to a truly individualized approach. Modern educational platforms can analyze each student's performance, identify their strengths and weaknesses, and adapt further content and tasks accordingly.¹

Platforms like Century Tech or Cognii use algorithms to provide the student with personalized recommendations and real-time feedback, adapting learning to their individual pace.¹⁵ If the system detects that a student is having trouble with a specific mathematical concept, it can automatically offer an additional explanation, a video, or an exercise focused precisely on that area. Conversely, for a student who masters the material without problems, the system can assign a more challenging task to deepen their knowledge.

This approach goes beyond the traditional differentiation that teachers have always tried to apply, but which was extremely difficult in practice with a larger number of students in the classroom. AI offers the possibility of scalable individualization. Virtual assistants and chatbots are also available 24/7, meaning a student can get help or an answer to their question at any time, even outside of school hours.⁶ This personalized approach is particularly beneficial for students with special educational needs, for whom AI can provide adaptive tools and targeted support tailored to their needs.¹

Using data to support learning (Learning analytics)

In addition to assessing individual students, AI can also analyze aggregated data on the performance of an entire class or school. These analyses, known as learning analytics, provide teachers with a valuable diagnostic tool for reflection and improvement of their own teaching.¹

For example, if an AI testing tool shows that 80% of students in a class made a mistake on the same question concerning a specific chemical formula, it is not just a signal of the students' lack of knowledge. It is, above all, valuable information for the teacher that their original explanation of this concept may not have been clear or illustrative enough.17 Based on this data, the teacher can adapt subsequent lessons, choose a different teaching approach, or focus on reviewing problematic topics.¹⁵

In this way, the role of the teacher also changes. From the position of an external evaluator who assigns grades, they become a data-informed diagnostician of the entire educational process. AI provides them with evidence on which they can make better pedagogical decisions. However, using this data also requires new competencies from teachers, especially basic data literacy and the ability to correctly interpret the pedagogical data that the systems provide.

 

A practical guide to tools and strategies for teachers

Moving from theory to practice requires knowledge of specific tools and strategies that teachers can realistically use in their work. This section offers an overview of available technologies and suggests specific methods for integrating AI into formative assessment and adapting assignments to the new reality.

Overview of key tools for assessment and feedback

There is a wide range of tools on the market with different focuses. For better orientation, they can be divided into several categories:

• Automated assessment tools: These include platforms like Quizizz and Kahoot!, which allow for the creation of interactive quizzes with immediate evaluation.⁵ More advanced systems like Gradescope can assess not only tests with closed-ended questions but also more complex tasks, such as mathematical calculations or short written answers, with AI intelligently grouping similar answers for faster correction by the teacher.¹⁵ Even common Google Forms can be turned into a powerful tool for automatic assessment with the help of AI extensions.⁵

• Feedback tools for written expression: Applications like Grammarly or LanguageTool provide instant feedback on grammar, style, and punctuation in Slovak and foreign languages. Students can use them to check their work before submission.⁵

• Originality and AI detection tools: Platforms like Turnitin are known for their ability to compare student work with an extensive database of sources and detect plagiarism. Newer versions also try to detect text generated by artificial intelligence, but it is important to approach their results with great caution, as their reliability is often low.⁷

• Comprehensive educational platforms: Some systems, such as Century Tech or SchoolHub.ai, offer integrated solutions that combine content creation, personalized learning, automated assessment, and data analysis for teachers on a single platform.¹⁵ 

The following table provides an overview of selected tools with specific examples of their use.

Table 1: Overview of selected AI tools for assessment

 

Strategies for using generative AI (ChatGPT, etc.) in formative assessment

Generative AI opens up new possibilities for formative assessment that go beyond simple correctness checking. Instead of worrying about how students use AI to answer questions, teachers should focus on how they can use AI to design better questions and tasks. AI thus becomes a partner in pedagogical design.

Here are some specific strategies:

• Co-creating assessment rubrics: The teacher can use AI together with students to generate a draft of an assessment rubric for an upcoming project. The subsequent discussion about the criteria and their adjustment leads students to a deeper understanding of what is expected of them and increases the transparency of the assessment.

• Generating case studies and scenarios: AI can create complex and relevant model situations for problem-based learning in seconds, which would otherwise take the teacher hours to prepare.¹⁰ For example, for a history lesson, AI can generate a fictional dialogue between two historical figures, which students then analyze.

• AI as a "critical friend": Students can be guided to ask AI for feedback on a draft of their work. For example, they can enter the prompt: "Act as an experienced literary critic and give me feedback on this analysis of a poem. Focus on the strength of my arguments and suggest where I could improve them."

• Differentiating tasks: The teacher can ask AI to adapt an assignment for different student levels. For example: "Create three variants of a word problem for calculating percentages. Let the first be basic, the second moderately difficult, and the third for advanced students."

Redesigning tasks and assessment methods in the AI era

The most effective response to the challenges that AI brings in the area of academic integrity is not its prohibition or the pursuit of perfect detection. It is the thoughtful redesign of what and how we assess. If AI can easily generate an essay on a given topic, then assessing such an essay as a final product loses its meaning.⁸

The logical consequence is that assessment must shift from the product to the process and focus on skills that AI (for now) cannot fully replace: critical thinking, creativity, complex problem-solving, and ethical reasoning.¹¹ This approach leads to the rise of metacognitive assessment - the assessment of thinking about thinking.

Practical strategies include:

• Assessing the process, not just the result: Teachers can assess different versions and edits of a document, thereby seeing how the student's work has evolved. Tools like Google Docs or Microsoft Word with the track changes feature are ideal for this.¹¹

• Oral defenses and discussions: Even if the work was partially created by AI, the student must be able to defend its content, explain their thought processes, and answer follow-up questions.

• Requiring reflection: Part of the assignment can be a mandatory reflection in which the student describes how they proceeded in creating the work, what tools (including AI) they used, how they verified the information obtained from AI, and how they critically evaluated its outputs.⁷

• Integrating AI into the assignment: Instead of banning it, AI can be directly incorporated into the task. Example assignment: "Use ChatGPT to summarize the main arguments for and against nuclear energy. Then, find two expert sources and write your own analysis in which you compare the AI's output with these sources and evaluate its accuracy and objectivity."¹¹

Tasks designed in this way not only reduce the risk of cheating but also develop key 21st-century competencies in students - the ability to critically work with information from various sources, including those generated by artificial intelligence.

Ethical challenges and responsible use in the school environment

Alongside its enormous potential, the integration of AI into assessment also brings serious ethical challenges that require attention and a proactive approach from teachers and school management. The most pressing issues include academic integrity, the risk of algorithmic bias, and the protection of sensitive student data.

 

Academic integrity and new forms of cheating

The availability of generative AI has created new and easy ways for students to cheat, for example, by generating entire essays or reports.⁸ The instinctive reaction of many schools is to try to introduce detection tools. However, these tools are often unreliable and can produce false positives, leading to wrongly accused students, a breakdown of trust between teacher and student, and unnecessary anxiety.⁷

The pursuit of perfect technological detection resembles an "arms race" that schools cannot win. AI models are constantly improving to make their outputs increasingly indistinguishable from human ones. A more sustainable and pedagogically meaningful solution is therefore to shift energy from repression to prevention. The key is to:

1. Establish clear rules: The school should have developed and clearly communicated rules that define under what conditions the use of AI in creating assignments is acceptable and when it is considered academic misconduct. These rules must be understandable to students and parents.²²

2. Focus on pedagogical solutions: As described in the previous chapter, the best defense is to redesign tasks to assess the process, critical thinking, and the student's unique contribution, making simple copying of AI output ineffective.¹¹

 

The risk of algorithmic bias and ensuring fairness

Algorithmic bias is one of the most serious, yet least visible, risks of AI in assessment. AI systems learn from vast amounts of data that often reflect existing societal prejudices - whether racial, gender, socioeconomic, or cultural. The algorithm not only reproduces these prejudices but can also amplify them.²³

Imagine, for example, an essay-grading system trained predominantly on texts written by native speakers from a specific cultural background. Such a system could systematically disadvantage students for whom the language of instruction is a second language, or who use different stylistic devices typical of their culture.²⁰ Algorithmic bias is thus not just a technical error; it functions as a form of hidden curriculum. When the system repeatedly gives lower scores to students from a certain group, it not only teaches them that their answer was "incorrect" but also implicitly communicates that their way of expression or thinking is less valuable. This can have a devastating impact on their self-confidence and motivation.

Furthermore, the General Data Protection Regulation (GDPR) explicitly prohibits discrimination based on automated processing of personal data.⁹ Teachers should also be aware that most current AI systems are designed to assess measurable knowledge and do not evaluate key skills such as collaboration, creativity, or social competencies.²⁰ The teacher's role is therefore to become an advocate for fairness and to ask critical questions of AI tool providers: "On what data was your model trained? How do you ensure that it does not discriminate against different groups of students?"

Protecting sensitive student data and GDPR compliance

The use of any online tool in school involves the processing of students' personal data—from their name and email address to the very content of their work, which speaks to their abilities, knowledge, and thoughts. This data is sensitive, and its processing is subject to the strict rules of the GDPR.⁹

A lack of trust from parents and students regarding the potential misuse of data is one of the main barriers to the wider adoption of AI in schools.¹ It is therefore essential that the teacher, when selecting and using AI tools, assumes the role of a responsible data controller. In practice, this means:

• Vetting the tool: Before a teacher recommends a tool to students, they should check its terms of use and privacy policy to ensure they comply with GDPR.18

• Being transparent: Teachers should openly communicate with students and their legal guardians about what tools are used in class, for what purpose, and what data is collected and processed.²²

• Teaching students digital hygiene: It is important to instruct students never to enter sensitive personal information about themselves or others into publicly available AI models (like the free version of ChatGPT).

• Respecting age restrictions: Many tools have minimum age limits for use. UNESCO recommends a general age limit of 13 for using generative AI and emphasizes the need to ensure data protection for minors.³

A responsible approach to data protection is a fundamental prerequisite for building trust and the ethical use of AI in education.

Conclusion: The teacher as an architect and ethical guide for education in the AI era

Artificial intelligence is not a passing trend but a transformative force that is changing not only the tools we use but also the very essence of our work. This chapter has shown that AI in assessment and feedback is neither a panacea nor a threat to be feared and avoided. It is a powerful tool whose impact depends entirely on how we handle it.

A summary of the key messages reveals a new, strengthened, and irreplaceable role for the teacher in the digital age. The teacher ceases to be merely a conveyor of knowledge and an assessor of results. They become:

• A critical evaluator of technologies: Capable of judging which tool is pedagogically appropriate, ethically safe, and aligned with educational goals.

• A designer of meaningful learning tasks: Who designs assignments that develop critical thinking, creativity, and collaboration - skills that AI cannot replace.

• An ethical guide: Who leads students towards the responsible, safe, and honest use of technology and helps them navigate the complex information environment.

This shift is fully in line with the human-centered approach advocated by UNESCO: technology should serve humanity, support its autonomy and dignity, not weaken it.³

Fear, ignorance, or bans are not sustainable strategies. The way forward is through curiosity, openness, and responsible experimentation. Teachers should be encouraged and supported to familiarize themselves with AI, test it on a small scale in their classes, and share their experiences and best practices with colleagues.²²

Ultimately, the most important competencies in the age of artificial intelligence remain human judgment, empathy, and the ability to build relationships. AI can grade a test and provide data, but it cannot inspire, motivate, and guide a young person on their journey of discovery. Artificial intelligence is a powerful tool, but its direction, meaning, and purpose will always be given by a wise, courageous, and ethically grounded educator.

 LADISLAV: ADD PRACTICAL EXAMPLE

Sources and recommended literature

1. UNESCO (2023). Guidance for generative AI in education and research. Paris: UNESCO. A key international document providing guidance for governments, schools, and educators on the ethical and effective implementation of generative AI in education.

2. UNESCO (2024). AI competency framework for teachers. Paris: UNESCO. A framework defining the key competencies that teachers should master in the age of artificial intelligence, from ethics to practical pedagogy.

3. Ministry of Education, Science, Research and Sport of the Slovak Republic. (2021). Ethical guidelines for teachers on the use of artificial intelligence and data in teaching and learning. A national document providing a framework for the ethical use of AI and data in Slovak education. (The reference to a specific document 20, although a European translation, is relevant to the context).

4. National Institute for Education and Youth (NIVAM). We recommend following the websites and publications of NIVAM, which regularly organizes training and publishes methodological materials on the topic of digital technologies and AI in education, for example, within the DiTEdu project.³⁰

5. AI for Children Initiative (AI dětem). A Czech non-profit initiative composed of educators and AI experts, which offers practical guides and materials for teachers, such as the ChatGPT guide for teachers.¹⁰

Works cited:

1. Umelá inteligencia (AI) prináša novú éru technológie na školách: Výhody a riziká, accessed on June 26, 2025, https://www.vysokahra.cz/umela-inteligencia-ai-prinasa-novu-eru-technologie-vyhody-a-rizika/

2. Umelá inteligencia v službách projektového vyučovania - Skolske.sk, accessed on June 26, 2025, https://www.skolske.sk/clanok/63424/umela-inteligencia-v-sluzbach-projektoveho-vyucovania

3. Guidance for generative AI in education and research, accessed on June 26, 2025, https://wp.table.media/wp-content/uploads/2023/09/386693eng.pdf

4. Artificial intelligence in education | UNESCO, accessed on June 26, 2025, https://www.unesco.org/en/digital-education/artificial-intelligence

5. Praktické metódy na využitie umelej inteligencie | Atlantis Center NEXT+, accessed on June 26, 2025, https://atlantiscenter.eu/praktick%C3%A9-met%C3%B3dy-na-vyu%C5%BEitie-umelej-inteligencie/

6. Žiaci sa učia dvakrát rýchlejšie: Učiteľov nahradila umelá inteligencia - Slovensko.AI, accessed on June 26, 2025, https://www.slovensko.ai/ziaci-sa-ucia-dvakrat-rychlejsie-ucitelov-nahradila-umela-inteligencia/

7. Umelá inteligencia rozdeľuje študentov a profesorov: Falošné obvinenia poškodzujú dôveru v školstvo, zistil výskum | Vosveteit.sk, accessed on June 26, 2025, https://vosveteit.zoznam.sk/umela-inteligencia-rozdeluje-studentov-a-profesorov-falosne-obvinenia-poskodzuju-doveru-v-skolstvo-zistil-vyskum/

8. Vplyv generatívnej AI na klasifikáciu univerzity: navigácia toku nadolových zručností, accessed on June 26, 2025, https://divmagic.com/sk/blog/generative-ais-impact-on-university-grading-navigating-the-down-skills-flux-ttgpiy

9. Etika a nástroje využitia AI v súvislosti s GDPR - GDPR.cz, accessed on June 26, 2025, https://www.gdpr.cz/etika-a-nastroje-vyuzitia-ai-v-suvislosti-s-gdpr

10. Ako využiť ChatGPT v edukačnom procese | direktor.sk, accessed on June 26, 2025, https://www.direktor.sk/sk/casopis/didaktika/ako-vyuzit-chatgpt-v-edukacnom-procese.m-1537.html

11. FGE | Umelá inteligencia v školstve - požehnanie alebo hrozba? - Future Generation Europe, accessed on June 26, 2025, https://futuregenerationeurope.eu/umela-inteligencia-v-skolstve-pozehnanie-alebo-hrozba/

12. Ako umelá inteligencia mení školstvo na Slovensku aj vo svete - Slovensko.AI, accessed on June 26, 2025, https://www.slovensko.ai/ako-umela-inteligencia-meni-skolstvo-na-slovensku-aj-vo-svete/

13. Ako AI pomáha s automatizáciou a správou vzdelávania - Unite.AI, accessed on June 26, 2025, https://www.unite.ai/sk/ako-ai-pom%C3%A1ha-s-automatiz%C3%A1ciou-a-riaden%C3%ADm-vzdel%C3%A1vania/

14. Umelá inteligencia ako súčasť vzdelávania, accessed on June 26, 2025, https://umelainteligencia.sk/umela-inteligencia-ako-sucast-vzdelavania/

15. 10 najlepších nástrojov AI pre vzdelávanie (jún 2025) (2025) – Unite ..., accessed on June 26, 2025, https://www.unite.ai/sk/10-najlep%C5%A1%C3%ADch-AI-n%C3%A1strojov-pre-vzdel%C3%A1vanie/

16. Nástroje umelej inteligencie pre dyslektikov a dysgrafikov | direktor.sk, accessed on June 26, 2025, https://www.direktor.sk/sk/casopis/didaktika/nastroje-umelej-inteligencie-pre-dyslektikov-a-dysgrafikov.m-1763.html

17. AI: Učiteľ v ére umelej inteligencie - Michal Gašparík-, accessed on June 26, 2025, https://www.michalgasparik.sk/2025/04/07/ai-ucitel-v-ere-umelej-inteligencie/

18. Bezplatná AI pre školy. - SchoolHub.ai, accessed on June 26, 2025, https://schoolhub.ai/sk-SK/get-started

19. Využitie ChatGPT pri tvorbe nástrojov formatívneho hodnotenia | direktor.sk, accessed on June 26, 2025, https://www.direktor.sk/sk/casopis/didaktika/vyuzitie-chatgpt-pri-tvorbe-nastrojov-formativneho-hodnotenia.m-2243.html

20. ETICKÉ USMERNENIA PRE PEDAGÓGOV TÝKAJÚCE SA POUŽÍVANIA UMELEJ INTELIGENCIE A ÚDAJOV PRI VÝUČBE A VZDELÁVANÍ, accessed on June 26, 2025, https://www.minedu.sk/data/att/c7a/32472.d707a3.pdf

21. Revolúcia na vysokých školách: Prežijú záverečné práce nástup umelej inteligencie?, accessed on June 26, 2025, https://www.trend.sk/nazory-a-komentare/revolucia-vysokych-skolach-preziju-zaverecne-prace-nastup-umelej-inteligencie

22. Pravidlá používania umelej inteligencie (AI) pre študentov a pedagógov PEVŠ, accessed on June 26, 2025, https://www.paneurouni.com/wp-content/uploads/2024/02/pravidla_pouzivania_ai-1.pdf

23. Aké sú hlavné etické výzvy pre ďalší vývoj modelov AI a ML? - Akadémia EITCA, accessed on June 26, 2025, https://sk.eitca.org/umel%C3%A1-inteligencia/eitc-ai-adl-pokro%C4%8Dil%C3%A9-hlbok%C3%A9-u%C4%8Denie/zodpovedn%C3%A1-inov%C3%A1cia/zodpovedn%C3%A1-inov%C3%A1cia-a-umel%C3%A1-inteligencia/ak%C3%A9-s%C3%BA-prim%C3%A1rne-etick%C3%A9-v%C3%BDzvy-pre-%C4%8Fal%C5%A1%C3%AD-v%C3%BDvoj-modelov-ai-a-ml/

24. Etické umělé inteligence: Rizika, odpovědnost a budoucnost rozhodování - Médium.cz, accessed on June 26, 2025, https://medium.seznam.cz/clanek/lukas-hrdlicka-eticke-umele-inteligence-rizika-odpovednost-a-budoucnost-rozhodovani-151624

25. Čo je zaujatosť umelej inteligencie? Príčiny, účinky a stratégie zmiernenia | SAP, accessed on June 26, 2025, https://www.sap.com/sk/resources/what-is-ai-bias

26. Etika a nástroje využití AI v souvislosti s GDPR - | TAYLLORCOX, accessed on June 26, 2025, https://www.tx.cz/blog/etika-a-nastroje-vyuzitia-ai-v-suvislosti-s-gdpr

27. Guidance for generative AI in education and research - UNESCO, accessed on June 26, 2025, https://www.unesco.org/en/articles/guidance-generative-ai-education-and-research

28. What you need to know about UNESCO's new AI competency frameworks for students and teachers, accessed on June 26, 2025, https://www.unesco.org/en/articles/what-you-need-know-about-unescos-new-ai-competency-frameworks-students-and-teachers

29. Summary of UNESCO AI Competency Framework for Teachers - CIDDL, accessed on June 26, 2025, https://ciddl.org/summary-of-unesco-ai-competency-framework-for-teachers/

30. Umelá inteligencia v škole – nástroje, ktoré ti pomôžu, nie ťa nahradia - NIVAM, accessed on June 26, 2025, https://nivam.sk/umela-inteligencia-v-skole-nastroje-ktore-ti-pomozu-nie-ta-nahradia/

31. AI competency framework for teachers | UNESCO, accessed on June 26, 2025, https://www.unesco.org/en/articles/ai-competency-framework-teachers

32. Kritické myslenie v ére umelej inteligencie (AI) – Národný inštitút vzdelávania a mládeže, accessed on June 26, 2025, https://nivam.sk/udalost/kriticke-myslenie-v-ere-umelej-inteligencie-ai/