top of page
Search

The AI Crisis in Education Today

Updated: 5 days ago

By Dakota Johnson

5/21/26


In classrooms and educational spaces across the country, artificial intelligence (AI) is transforming how students learn, how teachers teach, and how administrators manage. Yet most institutions lack a formal policy for such governance. Over the past semester, I’ve researched this policy gap, exploring how institutions, no matter the level, have failed to establish clear and consistent guidelines regarding generative AI. My research confirmed what most students and educators already feel: nobody has figured this out yet…and time is ticking. 


Whether you’re a student, educator, administrator, parent, or just a bystander, here’s what I’ve learned, and what you need to know: 

1. Institutions have failed to produce timely responses, even as AI increasingly becomes a part of our education today.

2. Existing policies lack uniformity; their variance across districts, schools, and universities weakens the credibility of success in regulation.

3. Students are left navigating AI’s uncertainty, with real consequences and administrative inaction.

4. Educators are left to solve a problem that wasn’t theirs to begin with. 


A Technology that Developed Quicker Than Policy


Since ChatGPT’s launch in November, 2022, generative AI and its tools have taken over learning environments: classrooms, libraries, and workspaces. According to UBS Financial Services, ChatGPT drew over 100 million users in just two months post-launch. ChatGPT’s offer of ease quickly wove itself into student life before society knew of its consequences.


Before alarms sounded in early 2023, it was already too late; AI had already established itself in our classrooms. My research drew heavily on Neil Selwyn, digital education specialist and professor at Monash University, whose work surveyed AI’s growing role in education today. 


He put it bluntly: educators were and are “positioned as having little control over the nature, pace, and direction of this technological change.” In his work, I noticed an alarming pattern: not only are institutions behind, but students are changing their work (and processes of understanding) to make better use of AI’s accommodations and education systems’ stronger demands. Selwyn found that students are “increasingly adapting to the algorithm,” further reshaping how they think, write, and complete assignments to mirror generative codes produced by AI. Yet, he also noted, AI “has no knowledge or understanding of what its output might mean.” In a sense, students are left pressured to use a tool that fails to simply comprehend its own production.


The Dependency Cycle 


For most students, AI dependency doesn’t start with cheating, but with a single search, prompt, and pure curiosity. Once clicking send, students are faced with instant answers. Over time, AI use becomes automatic—a crutch and ease for students who once produced original thoughts. Each prompt sent replaces an attempt at independent thinking.


What was once a shortcut becomes a crutch, and further, a routine. Eventually, students find themselves unable to perform because they’ve reached for AI without attempting to think for themselves, not out of laziness, but habit. This isn’t just an academic problem, but a behavioral one, where educational systems designed to protect students failed to see it coming.


Beyond grades and assignment performance, students are quietly eroded of something more personal: confidence, identity, and authorship. Research from a 2025 study at Cornell University by Kosmyna et al. found that students who completed tasks with AI assistance struggled to explain or recall their own work, whereas students who produced original, unassisted work were more confident in explaining their thought processes. For AI-assisted students, their work lay only in their prompting, as their authorship and thinking were separated under AI’s aid. In moments like these, students lose what researcher Hans Westerbeek calls “epistemic sovereignty”: the capacity to author knowledge from yourself, without reliance on predictive systems. To expand, as students allow AI to step in and think for them, they lose confidence in their own ability to produce it. Without growth in failure, which Westerbeek identifies as “epistemic weightlifting,” students lose the cognitive struggle that learning depends on. Similar to how lifting weights builds muscle, working through difficult problems, assignments, or projects builds neural pathways that are associated with memory, critical thinking, and reasoning, and further, executive functioning. For younger generations, exposed to intelligent systems capable of removing this critical step, consequences extend past habit, into developmental failures.  


What’s Actually Lost Here


Famous for his work on child development, psychologist Jean Piaget identified that critical thinking develops in an array of stages across childhood—and into adulthood, where each stage is crucially dependent on its predecessor. Although not the first stage of development, during the concrete operational stage (ages 7-11), logical thinking and problem-solving first take shape. Building on this foundation, the formal operational stage (ages 12+) deepens this development through independent decision-making and critical reasoning, both of which require sustained cognitive struggle and engagement to develop. 


Piaget's Stages of Cognitive Development
Piaget's Stages of Cognitive Development

Today, as AI reaches younger individuals, these critical stages are interrupted before students are able to develop the skills necessary to work through problems. Kosmyna et al.’s findings confirm this neurologically. As researchers measured brain activity across multiple groups (one of unaided individuals, individuals using search engines, and one using ChatGPT), unaided participants demonstrated the strongest neural connectivity, whereas ChatGPT users showed the weakest when measured with an EEG. In addition, when later asked to work without AI assistance, ChatGPT users struggled to re-engage with their work, demonstrating cognitive debt. Compared to older generations, where cognitive debt and struggle were inevitable, with the availability of technological assistance today, this struggle is increasingly optional. 


Every Institution Has a Different Answer, If at All


Academic integrity—and its processes—has existed long before AI. Institutions are constantly building and reworking their frameworks to protect it. When plagiarism took a high rise in the early ages of the Internet, honor codes, consequences, and detection software were set into place in response. Known by many, Turnitin, a plagiarism detector, has become the new norm; over 16,000 institutions worldwide use its software to uphold academic integrity today (CalMasters and The Markup). Because of communicated uniformity, these systems worked. Institutions agreed on the definition and worked together to solve it. Generative AI broke this consensus. Unlike previously known plagiarism, like copy-and-paste techniques, misciting statistics, or blatantly using someone else’s ideas, as of now, AI-generated work can’t be detected (at least reliably). Where a reported 58% of surveyed students admitted to using AI dishonestly in schoolwork and assignments, assuming truth exists within admitted faults, most institutions that these same students attend lack a formal policy to govern this AI use. 


The problem? An absence of uniform policy contradicts these institutions’ own AI implementation in our education systems today. For reference, by late 2023, about a year after ChatGPT’s launch, many universities had already begun integrating AI into teaching, administrative systems, and even student work; the majority of this integration lacked a strict framework. As a result, students sit quietly as their institutions adopt the same tools they’re penalized for using themselves. Beyond institution-to-institution differences, inconsistencies occur at the classroom level: some teachers welcome AI, others ban it, and some shift use from one assignment to the next. When no standard or policy exists, students are left with the message: it’s only wrong if it’s detectable, leaving integrity as optional. 


Nobody Told the Students


While institutions struggle to identify themselves amidst AI’s infiltration, students pay the price. Left without structured reform, students are forced to make judgment calls that aren’t theirs to begin with. Outside of the obvious frustration, this lack of solidarity is inequitable. In today’s educational age, resource access and allocation are strongly depicted by district funding. As a result, students from under-resourced schools are less likely to have proper AI literacy, highlighting the gaps in opportunities between institutions. That being said, with an absence of expectations, students with less guidance are most at risk. Without defined guidelines (and consequences), student dishonesty becomes a greater reflection of institutional failure. 


Teachers Are Alone

Also left without support, teachers and other educational instructors are left alone in deciding how to handle AI in classroom environments. Without administrative support, teachers are left unable to make consistent, enforceable decisions regarding AI use in their classrooms. Consequently, individual instructor choices vary with the possibility that two students down the hall from one another may operate under explicitly opposite expectations. This isn’t just a teacher problem, but a systems one—and worldwide. Until institutions take ownership and charge of this evolving technology, educators will continue to be burdened by this inconsistency.



The Bigger Picture


When institutions blindly adopt a technology without familiarizing themselves with its (potential) consequences, they fail to protect the students they’re responsible for. When students are punished for behavior modeled by these same institutions, and educators are left astray, we aren’t just behind, but broken. The systems we, as students, exist within today are struggling to keep up with the future they promise us. That’s worth paying attention to.


Sources:

Comments

GENCIVIC FULL LOGO

GENCIVIC - CLE

Empowering Cleveland-area youth to become the civic storytellers, changemakers, and leaders our democracy needs.

Email: gencivic.now@gmail.com

Subscribe to GenCivic Updates

bottom of page