Not everywhere, of course. But in enough places that the patterns are becoming clear.
From Policing to Pedagogy
The shift is real. In those early months, the instinct was to control, to detect, to prevent. But we are seeing assessment committees and teachers move beyond that initial reflex. As Zeivots et al. (2026) note in their recent study, institutions are now “balancing perspectives and priorities” in ways that acknowledge complexity rather than seeking simple solutions. The conversation has evolved from integrity violations to validity questions: What does authentic learning look like? What assessments truly measure what matters?
However, tensions remain. Guersenzvaig and Monett’s (2025) analysis reveals a “responsible realism,” where institutions acknowledge AI’s harms yet still promote its use, often shifting the burden of mitigation onto individual users. Their warning against “ethics washing” is sobering, but even the critics signal maturity: we’re past the honeymoon phase and into the hard work of critical engagement.
Rediscovering What Matters
Here’s the somewhat(?) unexpected gift: being forced to articulate what AI cannot do has clarified what we most value in education. When algorithms can generate competent text in seconds, we return to questions of meaning, context, judgment, and growth.
Dollinger and Nieminen (2026) reframe this disruption as opportunity. Rather than viewing GenAI merely as a threat, they propose it as “a catalyst for more equitable assessment that values collaboration and student agency over individual performance.” Their vision moves away from sorting students toward recognizing diverse ways of knowing.
This aligns with what Winstone et al. (2025) call the “ultimate feedback goal”: learning itself. Their manifesto emphasizes that our core purpose hasn’t changed even as our tools have. Teachers are navigating these “pedagogical tensions” in real time—not as problems to solve, but as productive sites of learning (Kohout-Diaz, 2026).
The Quiet Experimenters
Three years in, the real work isn’t happening in flashy press releases or top-down mandates. It’s happening in module teams, in course redesigns, in conversations between colleagues willing to try, fail, and try again.
Dilek et al. (2025) describe this as “critical co-discovery,” where educators position themselves not as experts delivering training, but as fellow learners exploring pedagogical and ethical implications alongside students. This approach is less about policy pronouncements and more about sustained, critical inquiry.
Evidence Emerges
We’re beginning to see what works—and what doesn’t. The OECD’s 2026 Digital Education Outlook documents innovative tools showing genuine promise. But the evidence also reveals pitfalls. Perkins et al. (2024, 2025) have evolved their work from a simple assessment scale to a critique of its misuse, warning against “performance theatre”—unenforceable bans and policies slapped onto unchanged assessments.
Substance over symbolism matters. We now know enough to identify that “AI-proofing” without redesign is a dead end.
Cracks and Foundations
“AI didn’t break education. It revealed the cracks that were already there.” is recited again and again. And for good reason —assessment practices overly focused on recall, learning outcomes that don’t distinguish genuine understanding from surface performance, and an overreliance on timed essays.
GenAI didn’t create these problems; it just made them impossible to ignore. And that is a good thing! Three years in, we’re examining our foundations. What capacities do students truly need when information is abundant? The answers aren’t simply about technology. They’re about humanity: meaning-making, ethical reasoning, and creative problem-solving in messy contexts.
So, yes
Slowly but steadily, it is happening. From assessment committees shifting focus from security to validity, to teachers engaging in co-discovery with students.
This isn’t a fast revolution we might have hoped for. It’s a quiet uprising of practitioners who refuse to let big tech dictate their values, who are learning—as their students must learn—to navigate ambiguity, to iterate, to keep the focus on what matters.
The sparks are there. Those who look carefully can see them growing.
Watch them shine.
References
- Dilek, Melis, Evrim Baran, and Ezequiel Aleman. “AI Literacy in Teacher Education: Empowering Educators Through Critical Co-Discovery.” Journal of Teacher Education, March 18, 2025, 00224871251325083. https://doi.org/10.1177/00224871251325083.
- Dollinger, Mollie, and Juuso Nieminen. “Reimagining Success and Failure: Equitable Assessment Practices in an Age of AI.” Journal of University Teaching and Learning Practice, ahead of print, January 26, 2026. https://doi.org/10.53761/20z06b11.
- Guersenzvaig, Ariel, and Dagmar Monett. Resisting Enchantment and Determinism: How to Critically Engage with AI University Guidelines. n.d. https://doi.org/10.5281/zenodo.18282338.
- Kohout-Diaz, Magdalena. “Making Sense of AI in Teacher Education: A Qualitative Study of Perceptions, Practices and Pedagogical Tensions.” Teaching and Teacher Education 171 (March 2026): 105342. https://doi.org/10.1016/j.tate.2025.105342.
- OECD. “OECD Digital Education Outlook 2026: Exploring Effective Uses of Generative AI in Education.” OECD Digital Education Outlook 2026 (January 2026). https://doi.org/10.1787/062a7394-en.
- Perkins, Mike, Leon Furze, Jasper Roe, and Jason MacVaugh. “The Artificial Intelligence Assessment Scale (AIAS): A Framework for Ethical Integration of Generative AI in Educational Assessment.” Journal of University Teaching and Learning Practice 21, no. 06 (2024): 06. https://doi.org/10.53761/q3azde36.
- Perkins, Mike, Jasper Roe, and Leon Furze. “How (Not) to Use the AI Assessment Scale.” Journal of Applied Learning & Teaching 8, no. 2 (2025). https://doi.org/10.37074/jalt.2025.8.2.15.
- Winstone, Naomi, Karen Gravett, Christy Noble, et al. “Manifesto for Feedback in the Age of Generative Artificial Intelligence.” Preprint, Figshare, 2025. https://doi.org/10.6084/M9.FIGSHARE.30195568.
- Zeivots, Sandris, Sue Wright, Lynne Harris, et al. “Walking the Tightrope of Quality Assessment: Balancing Perspectives and Priorities of Stakeholder Groups.” Studies in Higher Education, January 28, 2026. World. https://www.tandfonline.com/doi/abs/10.1080/03075079.2026.2619909.