Framing policies to enhance the education system in the age of AI
Despite several reforms by policymakers, investments, and sophisticated technologies, dissatisfaction with education systems persists. This review presents the challenges and provides a roadmap to enhance the education system in the age of AI. The author Yong Zhao highlights that education must move beyond borrowed solutions, narrow metrics, and automated fixes, and return to first principles. It is a call to resist reductionism and to recenter education on human growth, creativity, and responsibility.
For decades, education reform around the world has followed a familiar script. Policymakers look outward, borrow best practices, scale what works, and measure success through standardized outcomes. Despite waves of reform, mounting investments, and increasingly sophisticated technologies, dissatisfaction with education systems continues to grow. Student disengagement, teacher burnout, inequity, and a persistent mismatch between schooling and real-world demands raise a troubling question. What if the problem is not how education is improved, but how education is imagined?
This question has been at the center of Yong Zhao’s body of work published over the past 7 years in ECNU Review of Education. Rather than offering incremental solutions, Zhao’s scholarship challenges the foundations of modern schooling itself. Across issues ranging from international policy borrowing and large-scale assessment to pandemic disruption, technological transformation, and artificial intelligence, a consistent argument emerges. Education is constrained by an outdated paradigm, and meaningful improvement requires a fundamental shift in how learning, success, and human development are understood.
“The greatest obstacle to educational improvement is not a lack of solutions, but an inability to let go of old assumptions,” highlights Zhao.
When borrowing stops working
Zhao’s critique begins with a widely held assumption in global education reform, that systems can improve by importing policies and practices from high-performing countries. In his 2018 article, Shifting the education paradigm, Zhao argues that international borrowing has reached its limits, particularly in China. Policies designed for one cultural, economic, and historical context often fail to translate elsewhere, not because of poor implementation, but because education is not a modular technology that can be copied wholesale.
https://doi.org/10.30926/ecnuroe2018010105
Behind this argument lies a deeper concern. Reform efforts that focus on surface features such as curricula, standards, or instructional techniques leave intact the underlying assumptions about what education is for. Without questioning those assumptions, Zhao suggests, borrowing becomes a cycle of imitation without innovation.
Measuring what matters, and what does not
That concern becomes more pronounced in Zhao’s work on educational measurement. A cluster of articles published in 2019 examines the growing dominance of large-scale assessments and their unintended consequences. Together with collaborators, Zhao frames measurement as a “wicked problem.” What societies value most in education, including creativity, agency, ethical judgment, and collaboration, is often precisely what standardized tests cannot capture.
https://doi.org/10.1177/2096531119878965
https://doi.org/10.1177/2096531119868079
https://doi.org/10.1177/2096531119878964
Large-scale assessments, Zhao and his coauthors argue, do more than measure learning. They shape it. When test scores become proxies for quality, they narrow curricula, incentivize compliance over curiosity, and redefine success in reductive terms. The danger is not simply that tests are incomplete, but that they quietly replace educational purposes with measurable substitutes.
“When measurement becomes the goal, education loses sight of what truly matters.”
In this sense, Zhao’s critique extends beyond assessment policy. It exposes a broader pattern in modern schooling. Complex human capacities are reduced to technical indicators, and education becomes a system optimized for efficiency rather than meaning.
Crisis as revelation
The COVID-19 pandemic provided a dramatic stress test for these assumptions. In Tofu is not cheese, Zhao uses a simple metaphor to illustrate a profound insight. Moving schooling online without rethinking its structure is a case of mistaking imitation for transformation.
https://doi.org/10.1177/2096531120928082
The widespread frustration with emergency remote teaching, Zhao argues, did not stem from technology itself, but from attempts to preserve classroom routines that were never designed for digital environments. The pandemic did not create a crisis in education so much as reveal one. It exposed how deeply schooling relies on rigid schedules, physical co-presence, and standardized pacing, features that are poorly aligned with how learning actually occurs.
“The pandemic did not break education. It revealed how fragile our assumptions already were.”
Education in a technology-transformed world
As education systems gradually recovered from the pandemic, Zhao’s attention turned toward a more enduring transformation. Technology is no longer an instructional tool but a defining condition of modern life. In Time to rethink: Educating for a technology-transformed world, Zhao and his coauthor argue that treating technology as an add-on to existing models of schooling fundamentally misunderstands its impact.
https://doi.org/10.1177/20965311221076493
Under conditions where information is abundant, expertise is distributed, and learning extends far beyond school walls, long-standing distinctions begin to blur. The roles of teacher and learner, school and society, instruction and exploration are no longer fixed. Education must adapt not by adding more tools, but by redefining its purposes.
Artificial intelligence and the end of the grammar of schooling
The arrival of artificial intelligence intensifies these tensions. In a series of recent articles, Zhao advances his most far-reaching claim. AI does not simply challenge schooling practices. It undermines the very grammar on which modern education is built.
https://doi.org/10.1177/20965311241265124
https://doi.org/10.1177/20965311241296162
https://doi.org/10.1177/20965311251315204
Historically, schooling has been organized around fixed curricula, age-based progression, standardized assessment, and meritocratic sorting. AI destabilizes each of these elements. When machines can generate content, provide feedback, and personalize learning pathways, the rationale for uniform instruction and comparative ranking weakens.
Zhao and his collaborators argue that this moment calls for an ecological perspective on education. Learning must be understood as embedded in dynamic systems across time, space, and social relationships. In this view, the purpose of education shifts from producing individual competitors to cultivating human interdependence.
https://doi.org/10.1177/20965311251351988
“The future of education is not about winning against others, but about learning to live and work together in complex systems.”
AI also forces a reckoning within educational research itself. In examining the death and rebirth of research in the age of AI, Zhao and colleagues highlight both risks and possibilities. While algorithmic tools can trivialize inquiry and reinforce bias, they also open space for more human-centered, problem-driven, and ethically grounded forms of scholarship.
The latest work regarding this was published online in the journal ECNU Review of Education on August 19, 2025.
https://doi.org/10.1177/20965311251367053
Beyond reform
Taken together, Zhao’s work does not offer a blueprint for the next reform initiative. Instead, it poses a more demanding challenge. Education must move beyond borrowed solutions, narrow metrics, and automated fixes, and return to first principles.
Zhao’s critique is not anti-technology, anti-assessment, or anti-school. It is a call to resist reductionism and to recenter education on human growth, creativity, and responsibility. In an era defined by uncertainty, rapid technological change, and global interdependence, this reimagining is not optional.
Education, Zhao suggests, must help learners navigate, shape, and sustain the worlds they inhabit.
More about Zhao’s work can be found at https://zhaolearning.com/.
Reference
Title of original paper: The Death and Rebirth of Research in Education in the Age of AI: Problems and Promises
Journal: ECNU Review of Education
DOI: https://doi.org/10.1177/20965311251367053
Funding information
The authors received no financial support for the research, authorship, and/or publication of this article.