What does learning look like in the age of AI?

This is part of Kami’s ongoing series of executive briefings for education leaders. Each month, we’ll discuss solutions to a top-of-mind issue and share actionable resources and inspiration for leading your team. Be sure to subscribe to have this free resource delivered to your inbox.

Summary

• How to use authentic assignments and assessments as a healthier, pedagogically sound response to AI and academic integrity
• Why it’s important  to move beyond cycles of AI-generated content, surveillance, and discipline
• How-to guide and strategies for integrating authentic learning in schools

In the age of AI in education, how might educators move beyond a cycle in which students use AI to generate answers and teachers try unsuccessfully to catch them, and toward a more human-centered approach focused on process and application of knowledge? It turns out that AI exposed the learning design flaws that predated it. In this brief, we’ll examine the root causes of low engagement and academic dishonesty and offer actionable research-based solutions and practical classroom applications.

The Challenge

Teaching and learning used to be pretty straightforward: students would predictably show up to school ready to learn. Teachers would present a lecture or lab, assign readings, give a quiz or assign a paper, and students would earn grades that supposedly reflected their depth of knowledge.

Since the COVID-19 pandemic, students’ willingness to to show up physically and intellectually has continued to be a problem. But six years out, student test scores stubbornly remain far below pre-pandemic levels (Miller et al., 2026), and schools continue to struggle with attendance, signaling that something more systemic is at the heart of these challenges that affect learning outcomes and school district budgets.

In addition to chronic absenteeism, students now have access to AI both in and out of school. From searching for content to analyzing data, writing essays, and even solving math problems, the need for teachers to assess students in ways that are no longer just about the outcome or result is growing. And student engagement, a traditional indicator of student attendance, can no longer be just about seat time and compliance. Educators need assessment strategies that measure deeper, more authentic learning, and districts need to make a stronger case to students and parents about why showing up to school matters. 

These two challenges, engagement/attendance and meaningful assessment, go hand in hand. (Anderman & Koenka 2017). This brief is a guide for education leaders seeking effective, scalable, and practical solutions to these challenges.

According to a 2024 Gallup study, between a quarter and a half of K–12 students no longer find school interesting or relevant to their strengths (Gallup, 2024a). Student test scores have been in decline long before the COVID-19 pandemic as well (Miller et al., 2026). As conversations around AI in the classroom have been at the forefront of many district board meetings, there has been research that shows that the overwhelming majority of teachers and students are using AI within a single school year (Pew Research Center, 2026).

Under that pressure, district school boards and leaders often have one of two responses. The first is the detection arms race: buy plagiarism detection software, detect students who are academically dishonest  , and hope to restore the learning process. The second is a blanket ban:prohibit screen time, remove devices, and return to an educational model that predates technology. But new research finds that neither approach works (Klein, 2026).

To truly address these issues, education leaders need to treat the underlying causes of both chronic absenteeism and AI-related plagiarism, not just the symptoms. Learning artifacts such as worksheets or problem sets often measure only student compliance or time spent working, rather than the thinking required to develop an idea over time. Traditional assessment measurements also don’t allow students to develop and produce learning products based on original ideas. This also makes learning outcomes more resistant to AI generation.

What the research says about authentic learning

Student engagement leads to better attendance. Researchers find that engagement doesn’t precede attendance, and that to truly boost student interest in school, student-centered pedagogy is particularly effective. (Trani & Hart, 2023)

AI Plagiarism-detection software is flawed and biased. Independent testing of AI-text detectors found them unreliable, routinely flagging human writing as machine-generated, which makes them unsound as evidence of student plagiarism (Weber-Wulff et al., 2023 [VERIFY]). An integrity strategy built on detection manufactures false accusations and erodes the trust on which learning depends.

It may be the assignment, not AI. When a task can be completed without engaging in the learning it was meant to produce, generative AI simply makes the issue apparent (Kofinas et al., 2025). The productive response is to provide professional learning for help teachers redesign assignments so that they require personal context, real-time demonstration, documented process, and an authentic audience. In these assignments, authentic learning assessments become simultaneously harder to plagiarize and better aligned to durable learning (Ashwin et al., 2025).

School culture is a big factor in attendance and engagement. Qualitative action research underscores that positive peer interactions, a sense of belonging, and explicit teacher support are among the strongest self-reported motivators for students to maintain consistent attendance. (Smith, 2025)

Authentic learning raises both engagement and test scores. Educators often face the false choice between rigor and engagement. In a cluster-randomized controlled trial across five urban districts, students in project-based Advanced Placement courses earned credit-qualifying exam scores 8 percentage points higher than peers in lecture-based versions after one year, roughly 10 points higher after two, with gains holding for both lower- and higher-income students (Saavedra et al., 2022). Other studies reached the same conclusion: project-based learning improves academic achievement and higher-order thinking over traditional instruction (Zhang & Ma, 2023). These kinds of assignments are not a trade-off against test performance, but rather a way to improve it.

Implications for leaders

Treat integrity as a design problem, not a surveillance problem. The goal is not to “catch” student AI use; the learning goal shifts to make student thinking visible. Assignments built on originality, personal connection, and purpose prove that learning is happening by design rather than by fear of  detection. Replacing cultures of mistrust with cultures of integrity is also vital for student mental well-being, which has a measurable impact on academic achievement.  (Hashemzadeh, 2026) 

Stop framing test scores and authentic learning as opposites. The evidence proves otherwise: the most engaging, AI-resistant student assignments tend to be those that raise achievement (Saavedra et al., 2022; Zhang & Ma, 2023). Lead with visible outcomes, in which student agency, durable skills (such as critical thinking, collaboration, empathy, and citizenship) , and community connection are central to the curriculum and its assessments. This can be especially helpful for school boards and stakeholders anxious about student AI accountability.

Not all screen time is the same. Intentional use of tech provides digital and AI literacy, as well as mandated accessibility features. Students can’t learn AI and digital literacy skills by reading about them; they need to engage with these tools in safe, intentional ways that are aligned to curriculum and standards. Federal law also mandates that learning materials have accessibility features, many of which are impossible to provide without educational technology tools and even some AI tools. When making the case to parents, caregivers, and school boards, be clear about your obligation to provide equitable learning experiences for all students. 

Define what counts as evidence of learning, then live it. For many parents and teachers, evidence of learning has meant graded essays, tests, or class presentations. The shift to AI-resistant assignments is one that builds on these traditional learning models to create authentic learning experiences and artifacts. Learning artifacts that are created for an audience beyond the classroom, or that can be collected into student learning portfolios. Many states and school districts are pivoting to a Portrait of a Graduate model, in which these ideals are integrated into the curriculum and district culture.

Implementation considerations

Binding constraints are often the educator’s mindset and teacher capacity. Most teachers were never trained in assessment design, and asking them to rebuild learning assessments without professional learning or support may lead to  resistance and burnout. 

Sustained, job-embedded professional learning rather than one-day professional development (coaching, teacher-led collaboration, peer observation) is one of the highest-value levers that can lead to sustained change when it comes to combatting teacher burnout. Assisting educators with building authentic assessments into their instructional design can reduce it so that it is not seen as an add-on. 

Families also need clear messaging around authentic learning and decreased reliance on test scores as the only metrics for success. For many parents, GPA and test scores are cultural reference points for judging not only their students’ achievement and ranking but also a district’s overall performance and value. Creating genuine understanding through community outreach and forums will help with buy in. 

Beware of initiative fatigue. To avoid losing teacher buy-in before day one, leaders must use this new approach to replace outdated practices rather than just add to them. Here are evidence-based ways to navigate this.

Recommended next steps

The test-score pressure is not going away, and pretending otherwise can cost leaders credibility. The best strategy is to acknowledge this reality directly and show how more authentic, cognitively demanding work, not test prep, is, in fact, the fastest route to durable score gains (Saavedra et al., 2022; Zhang & Ma, 2023). 

Here are some strategies to make authentic learning part of the culture of your district:

Name what counts.  Put your definition of real learning in writing for teachers, stakeholders, and families. True learning means a student can apply knowledge independently, use tools effectively, and exercise good judgment. Be explicit about these goals and show exactly how project-based or authentic learning experiences meet your district’s academic standards.

Pilot Before Scaling.  Don’t issue assessment redesign mandates. Instead, start with a small cohort of willing teachers in a single, high-stakes course. Have them redesign a few assignments or one full unit into an authentic learning project. Set clear, measurable goals using the AP, state, or benchmark data you already track. This speaks the language your parents, students, and school board understand and allows for data-driven comparisons of student outcomes.

Pair every restriction with a redesign. If it is decided that access to technology should be limited, decide beforehand how students will handle research, data analysis, and authentic audiences for their work..  To maximize authentic learning, conduct reflections from both teachers and students to maintain initiative credibility.

Build a tiered evidence system.  Use standardized tests for external accountability, but allow classroom-based performance tasks to drive actual learning and standards for measuring student learning and growth. Trying to force one test or grading tool to do both jobs satisfies neither (Center for American Progress, 2023).

Measure agency and responsibility, not compliance. What a district measures is what it will get as outcomes. Ask teachers to assess what the district has determined are your values and expected learning outcomes. If your district values independent thinking and problem-solving, grade those skills; don’t default to tracking compliance or arbitrary grading practices. As active learning environments may look and assess differently from traditional classrooms, expect a period of transition. Give teachers the time and coaching they need to adapt to these new expectations.

Citations

Allcott, H. (2026). The Effects of School Phone Bans: National Evidence from Lockable Pouches (Working Paper No. w35132). National Bureau of Economic Research. https://www.nber.org/papers/w35132

Anderman, E. M., & Koenka, A. C. (2017). The relation between academic motivation and cheating. Theory Into Practice, 56(2), 95–102. https://doi.org/10.1080/00405841.2017.1308172 

Ashwin, P., et al. (2025). AI-based digital cheating at university, and the case for new ethical pedagogies. Journal of Academic Ethics. https://doi.org/10.1007/s10805-025-09642-y

Center for American Progress. (2023, October). Future of testing in education: Effective and equitable assessment systems. https://www.americanprogress.org/article/future-testing-education-effective-equitable-assessment-systems/

Fortune Business Insights. (2026, May 4). Anti-plagiarism for the education sector market size, share & COVID-19 impact analysis, by type (text mining, and integrated customization service), by application (students, and teachers), and regional forecast, 2020-2027 (Report ID: FBI104552). https://www.fortunebusinessinsights.com/anti-plagiarism-for-the-education-sector-market-104552

Gallup. (2024a, August). K-12 schools struggle to engage Gen Z students. https://news.gallup.com/poll/648896/schools-struggle-engage-gen-students.aspx

Hashemzadeh, K. (2026, January 26). New study provides evidence that social-emotional learning programs improve academic performance. USC Rossier School of Education. https://rossier.usc.edu/news-insights/news/2026/january/new-study-provides-evidence-social-emotional-learning-programs-improve-academic-performance

Kaplunov, E., & Ntasioti, A. (2025, September 22). Is AI making students less critical thinkers? Advance HE. https://www.advance-he.ac.uk/news-and-views/ai-making-students-less-critical-thinkers

Klein, A. (2026, May 4). Do student cellphone bans improve academic achievement? Education Week. https://www.edweek.org/technology/do-student-cellphone-bans-improve-academic-achievement/2026/05

Klein, A., & Langreo, L. (2026, April 20). The ed-tech backlash is here. What it means for schools. Education Week. https://www.edweek.org/technology/the-ed-tech-backlash-is-here-what-it-means-for-schools/2026/04

Kofinas, A., et al. (2025). The impact of generative AI on academic integrity of authentic assessments within a higher education context. British Journal of Educational Technology. https://doi.org/10.1111/bjet.13585

Miller, C. C., Paris, F., & Mervosh, S. (2026, May 13). Why U.S. test scores are in a “generation-long decline.” The New York Times. https://www.nytimes.com/2026/05/13/upshot/test-scores-school-districts-us.html

National Assessment Governing Board. (2025, January). The Nation’s Report Card: Declines in reading, some progress in 4th grade math [Press release]. https://www.nagb.gov/news-and-events/news-releases/2025/nations-report-card-decline-in-reading-progress-in-math.html

Pew Research Center. (2026, February). How teens use and view AI. https://www.pewresearch.org/internet/2026/02/24/how-teens-use-and-view-ai/

Saavedra, A. R., Lock Morgan, K., Liu, Y., Garland, M. W., Rapaport, A., Hu, A., Hoepfner, D., & Haderlein, S. K. (2022). The impact of project-based learning on AP exam performance. Educational Evaluation and Policy Analysis, 44(4), 656–680. https://doi.org/10.3102/01623737221087474

Smith, Q. (2025). The Impact of Students’ Attendance on Their Academic Success: A YPAR Study (Doctoral dissertation). Scholar Commons. https://scholarcommons.sc.edu/etd/8475/

Trani, J.-F., & Hart, R. (2023). Student engagement and attendance are central mechanisms interacting with inclusive and equitable quality education: Evidence from Afghanistan and Pakistan. PLOS ONE, 18(12), e0290456. https://doi.org/10.1371/journal.pone.0290456

Weber-Wulff, D., Anohina-Naumeca, A., Bjelobaba, S., … Waddington, L. (2023). Testing of detection tools for AI-generated text. International Journal for Educational Integrity, 19(1), Article 26. https://doi.org/10.1007/s40979-023-00146-z

Zhang, L., & Ma, Y. (2023). A study of the impact of project-based learning on student learning effects: A meta-analysis study. Frontiers in Psychology, 14, 1202728. https://doi.org/10.3389/fpsyg.2023.1202728

Get our complete brief for education leaders delivered to your inbox every month.

Main illustration by Michael Hernandez