Mind the Grade: Authentic Assessment in the Age of AI

Artificial intelligence is evolving rapidly, leaving education in a state of flux, but what if the most important move is to pause and think?

AI’s uncertainty is real, and its accelerating presence in schools has left educators feeling overwhelmed.  This moment calls for an honest inquiry into the critical underpinnings of AI’s pervasive impact on the education sector, particularly on classroom instruction and assessment, for consideration of who gains from AI, and more importantly, at what cost?

Walk into almost any school today, and you’ll hear a common question among educators: How do we know if this is actually students’ work? This query has a prolonged history in education, asked long before the digital age, reignited when digital tools entered classrooms, and is now being asked again with a new sense of urgency.  Perhaps reframing the question is more useful than chasing it. It asks teachers to courageously and critically consider all facets of AI and its current role in education.  Rather than how do we catch students using AI, we might ask: how do we design learning experiences so authentic that AI use becomes inconsequential?

Start With Why

With the dawn of the digital age came one of the most important lessons for the twenty-first-century classroom: every (digital) tool needs a purpose, and understanding the purpose behind each tool matters.  Before any technology enters a classroom, educators need to ask the question: Why this tool, for this student?

The digital age has shown that the right technology, when deployed thoughtfully and intentionally, can remove barriers that once seemed impossible. Accessible learning has enabled students to offload cognitive demands that once overshadowed deeper learning, helping to build inclusive spaces and meaningful student participation that had previously been out of reach. The emergence of today’s AI technology demonstrates the potential to enable new forms of teaching and to enhance student learning experiences (UNESCO, 2024). When technology serves genuine accessibility needs, its use should not only be permitted but actively supported and protected. However, as appealing as AI can be, it also poses significant risks beyond accessibility. In classrooms, the use and overreliance on AI can be reduced to automated tasks, weakening the relational dynamics among students and teachers (UNESCO, 2024). AI also poses broader risks, including violations of data privacy, intensification of the climate crisis, perpetuation of inequities, and threats to human agency (UNESCO, 2024). Recognizing the moral dilemmas of AI means placing ethics at the centre of AI discourse (European Commission, 2025).

Where Are We Now?

Artificial intelligence did not arrive overnight. It has been subtly and steadily entering our daily lives for years in the form of reactive (e.g., virtual assistants) and predictive (e.g., predictive text) entities.  Today’s generative AI presents a paradox that carries both extraordinary promise and genuine risk. On the one hand, it offers insight to help cultivate personalized learning experiences for students, and it continues to change the educational landscape on many levels.  But it also arrives in a landscape already contending with the harms of smartphone use and excessive screen time among young children and youth.  As educators navigate the challenges of classroom assessment, this well-documented evidence has the potential to offer something valuable.  It serves as a guide, a framework grounded in evidence-informed decision-making that could serve us well when discerning the use of AI in education.  As AI platforms continue to roll out across school boards, it is worth asking: Are these decisions driven by a genuine commitment to student learning or by the pressure to appear current or ‘innovative’?   Is the privacy and safety of students and teachers placed at the center of AI discourse?  What is the cost of chasing the AI perpetual wheel?  Ultimately, in this massive enterprise of Ed-tech, who does AI fundamentally serve? Unlike any other technology seen before,  AI is not a tool, but an agent (Harari, 2024).  This should trigger a collective pause to ensure decisions are made in the best interest of all.  Fundamentally, AI does not change our purpose, our duty of care, and our steadfast duty to educate (OASBO and ECNO, 2025).

Authentic Assessment

When the crew of Artemis II returned from their space journey, they shared how no amount of quantitative data could truly capture what the four of them had experienced together.  The tangible data would only be able to tell part of the story, but the rest of the story—the nuance of human connection— belonged to the four of them.  Classroom culture, though not as astronomical in scale, is sprinkled with such moments. The expression on a student’s face when something finally clicks. The quiet conversation after class that changes how a teacher understands a student, both their struggles and triumphs. The laughter that erupts when an analogy lands perfectly, or the silence that falls when a question opens something unexpected. These are not incidental to learning. They are the learning, in its rawest, intangible and irreducible form.

Leveraging the possibilities of AI with its implications is complex.  The current Growing Success (2010) framework in Ontario asks teachers to triangulate evidence of student learning through three data points: conversations, observations, and products. Through triangulation, a grade or mark is not merely a score but rather a culmination of an ongoing, relational process where assessment information is gathered both in the moment and as part of a broader, iterative exercise. The classroom teacher, for instance, may consider how a student explains their thinking aloud, or in the hesitation before an answer.  Teachers may consider how a student demonstrates their learning in written, auditory, and/or visual form, or in the questions they might ask along the way.  These are observable moments that happen in the room, in real time, between and among teachers and students.  This is precisely why homework, under current Ontario policy, cannot be formally evaluated.  It isn’t witnessed, and its authenticity cannot be confirmed by the classroom teacher when determining a grade. This reasoning holds just as firmly for AI-assisted work, or any work, produced outside the classroom. The solution may not be to increase surveillance but rather to bring more learning back into the observable, humanized space of the classroom itself. It means focusing on what AI cannot replicate: human presence.  An authentic understanding of students’ thinking and learning reveals itself in conversation and observation in ways AI cannot sustain under scrutiny. The goal, then, is not to ‘catch’ a student, but to create the conditions where their real learning can be revealed honestly and safely.  AI can process language, generate text, identify patterns, and eventually (perhaps?) personalize learning in infinite ways. But it cannot be present. It cannot notice and wonder in the full, embodied sense that a teacher does when they are truly with their students. Teachers cannot be replaced by technology as human flourishing should remain at the center of the educational experience (UNESCO, 2024).  The capacity to be a witness to another person’s growth remains uniquely human.

What Educators Can Do

Today’s classrooms reflect a complex terrain.  Educators are in the midst of navigating many unknowns while contemplating the pros and cons of AI, on the fly.  UNESCO’s AI Competency Framework for Teachers (2024) has been developed to support systems of education with the principles of protecting teachers’ rights, enhancing human agency, and promoting sustainability.  With a human-led approach, it provides a comprehensive framework that outlines competencies and considerations as a starting point for professional learning and classroom practices.  With the rise of AI detectors, the return of fully analog approaches, and the growing sentiment to “ban” various tools, books (sigh…) and platforms, teachers are called to critically discern how to meet this moment.

1. Professional Learning 

• Engage in AI discourse to stay informed, ask critical questions, and build trust throughout the learning process. Professional learning, as a vehicle for capacity building, should be intentionally designed and facilitated at both a system and local level to enable lifelong learning.  To support this multi-levelled approach, teachers need to first understand how AI systems are designed and how AI models work, in order to protect human agency, linguistic and cultural diversity, and Indigenous knowledge (UNESCO, 2024).
• Review current AI Guidelines both locally and globally.  Inquire about any existing school board guidelines and safety guardrails in place designed to protect the privacy and well-being of students and teachers.  AI guidelines must be iterative, living documents that are constantly revisited, reviewed, and rewritten to reflect diverse perspectives for an evolving entity that itself refuses to stand still.  There needs to be a critical dismantling of the seductive myth that tech innovation is inherently harmless.
• Revisit, reflect and discuss current provincial and school board assessment and evaluation policies.  Consider the negotiables and non-negotiables when students are invited to demonstrate their learning, and how teachers exercise their professional judgment to determine a grade.  When navigating decisions about AI, placing safety, accessibility, and ethical discernment at the core of classroom assessment practice is fundamental to a human-led approach.  AI tools should never be designed to replace the legitimate responsibility and accountability of teachers, and classrooms should remain as spaces where human and social values prevail (UNESCO, 2024).

2. Classroom Practices   

• Feedback Over Grades: Formative feedback moves thinking and learning forward by developing and sustaining intrinsic motivation over extrinsic validation (Ryan & Deci, 2000).  Planning and assessment are at the heart of the teaching profession.  By shifting the emphasis on formative assessment, in real time, it not only improves achievement but deepens the relational dynamics in the classroom.
• Questioning Strategy: In Ontario, assessment and evaluation are based on the achievement of the overall expectations, or the ‘big ideas.’  Engage students with thought-provoking questions that cannot be directly answered through the use of any digital tool.  Asking critical, open-ended questions can provoke deep thinking and learning that directly connects to these expectations.
• Talk Structures/Conferencing: Elicit presence and engagement through real-time conversations as assessment for, as, and of This elevates and authenticates professional judgement about student work, thinking and learning when triangulating data.
• Content Selection: Skills-based curriculum allows for a variety of diverse texts to be selected and explored. Intentionally select AI content, articles, and evolving guidelines, etc., as texts for This supports students’ awareness and critical thinking skills about current research, bias, ethics, and the environmental impact of AI.

There is much at stake in education as the world grapples with the sweeping forces of AI, which are already shaping the future of learning, work and careers.  Educators want to honestly reckon with the possibilities and repercussions—at every scale—and thoughtfully consider how AI impacts the texture of daily classroom life. Professor Geoffrey Hinton (aka “The Godfather of AI”) recently shared his thoughts in response to AI and the next generation:

“Get yourselves a good liberal education that teaches you to think. Learn some STEM because it would be good to understand what is happening… but you should also learn to think. The people who are going to survive longest are the people who are going to be able to think.”

Moving forward in the age of AI does not mean moving fast, but rather slowing down to build the conditions in classrooms and schools, in education policy, and in professional culture where deep, critical thinking can happen. Students need to be valued for the minds they develop, not for the products they produce, and teachers are to be trusted for how they critically question and exercise their professional judgement, and not for the extent to which they simply abide by AI influence, fallacy, and power.  Authentic assessment is about humaneness in learning, not algorithmic evaluation—students are “more than a score” (Hagopian, 2014).  Ultimately, the classroom should remain a place of possibilities where the question is never simply what can AI do? But rather, what do we want learning to be?

AI will keep evolving, and so too should the guardrails, but what must remain constant is the irreplaceable experience of being fully present.  This is worth protecting.  The most important thing AI has revealed about education may not be what it can automate, but what it can’t.

That’s worth thinking about.

Reflection Questions

System Leaders:

Review your school board’s current and/or future AI guidelines and consider the following:

1. Whose voices are represented? Whose voices are missing?
2. How are multiple perspectives reflected in the guidelines, in ways that support educators’ and students’ discernment in their use of AI?
3. How are the AI guidelines “living,” human-led documents? How will they be reviewed and revised to ensure that the representation of diverse voices and perspectives, and the safety and well-being of all students and staff, are honoured?
4. How will relevant and evidence-informed data on AI guide future iterations of the AI guidelines?

Classroom Educators:

1. What confirms your thinking about AI in relation to your current classroom assessment practices? What challenges your thinking?
2. What might you stop-start-continue doing in light of new learning about AI and what you’re noticing in the classroom?
3. How might you weave critical AI literacy into your professional practice and discourse with students, colleagues, caregivers and system leaders?
4. How will you exercise your professional judgement in student assessment while navigating the AI landscape?
5. What questions do you have? What supports and/or professional learning do you need to deepen your thinking about AI?

References

European Commission. (2026). Guidelines on the ethical use of artificial intelligence and data in teaching and learning for educators (Revised ed.). Publications Office of the European Union. https://doi.org/10.2766/7967834

Hagopian, J. (2014). More than a score: The new uprising against standardised testing. Haymarket Books.

Harari, Y. N. (2024). Nexus: A brief history of information networks from the Stone Age to AI. Random House.

Mac, A. (2026, March 31). The AmberMac Show Ep059: Godfather of AI Geoffrey Hinton on chatbots, AGI & Putin (Part 2). AmberMac. The AmberMac Show Ep059: Godfather of AI Geoffrey Hinton on Chatbots, AGI & Putin  (Part 2)

Miao, F., & Cukurova, M. (2024). AI competency framework for teachers. UNESCO. https://unesdoc.unesco.org/ark:/48223/pf0000391104

Ontario Generative AI Innovation Working Group. (2025). Guidelines for responsible use of generative artificial intelligence (v1.3). Educational Computing Network of Ontario (ECNO).https://www.ecno.org/wp-content/uploads/2025/04/English-Generative-AI-Guidelines-v1.3.pdf

Ontario Ministry of Education. (2010). Growing success: Assessment, evaluation, and reporting in Ontario schools, covering grades 1 to 12. Queen’s Printer for Ontario.

Ryan, R. M., & Deci, E. L. (2000). Self-determination theory and the facilitation of intrinsic motivation, social development, and well-being. American Psychologist, 55(1), 68-78. https://doi.org/10.1037/0003-066X.55.1.68

Photo supplied by author.

Meet the Expert(s)

Patt Olivieri

K-12 Literacy Resource Teacher, Toronto Catholic District School Board & Doctoral Student, Ontario Tech University

Patt Olivieri has been a K–12 educator for almost thirty years. As a former Early Childhood Educator, she brings a deep curiosity about how students learn, grounded in an inquiry-based approach through pedagogical documentation. Patt has held a variety of roles in elementary and secondary education, including Special Education, Intermediate/Senior English, the Care & Treatment Program, and Assistive Technology. She currently works as a Literacy Resource Teacher for the Toronto Catholic District School Board (TCDSB), is an Additional Qualification instructor for the Ontario English Catholic Teachers' Association (OECTA) and York University, and is pursuing doctoral studies in ungrading assessment practices at Ontario Tech University.

Read More