AI Aced Your State’s Gen Ed. Now What?

Time to Move Upstream

It’s Fall 2026. A first-year college student planning to study environmental science sits in her academic advisor’s office at State University, reviewing her general education requirements. Like most of her peers, she’s been using ChatGPT since tenth grade for homework help, essay drafting, math problem-solving, and research synthesis. Now her advisor is explaining that she needs 36 credits across writing, quantitative reasoning, scientific literacy, and social sciences to meet the state’s general education mandate. The student politely listens, puzzled: why spend a year and a half taking courses to develop skills that AI has been handling for her for years? The advisor mentions critical thinking and information literacy outcomes, but the student knows that current AI systems can analyze arguments, evaluate sources, and synthesize research more thoroughly than most humans. AI helped her write her application essay on 100 years of core samples from Greenland glaciers, for goodness sakes, from the quiet of her midwestern suburban home.

Here’s an urgent reality: in every state, higher education is floundering as AI becomes the world’s primary platform for accessing and demonstrating general knowledge. Standard state-mandated general education requirements focus on two elements: mastery of general knowledge across disciplines and demonstration of core academic skills. AI platforms now excel at both. AI already delivers 90% of the standard requirements across state systems: demonstrating written communication, quantitative reasoning, critical analysis, and information literacy. Large language models can write clear prose across disciplines, solve complex mathematical problems, analyze arguments for logical flaws, and synthesize information from multiple sources. They can explain scientific concepts, analyze historical events, and engage with philosophical ideas at a level that would pass nearly every gen ed assessment. They can make every student workforce-ready.

Universities must move upstream to where knowledge is created and validated. This means fundamentally reconceptualizing gen ed, if there is any future role for it at all. Instead of focusing on transmitting established knowledge or practicing basic academic skills, universities must teach each student how scholarly communities discover, validate, and create new knowledge. This work is inherently individual and cannot be replicated by AI. Requiring students to take general courses to demonstrate competencies that AI can replicate serves no educational purpose. The university’s role in the AI era is teaching students how to participate individually in knowledge creation, not just knowledge consumption.

This shift isn’t speculative, it’s measurable against current assessment standards. The standardized rubrics used to assess general education outcomes, from AAC&U VALUE rubrics to state-specific frameworks, list specific competencies that AI systems already demonstrate. Take writing outcomes: “clear thesis,” “supporting evidence,” “logical organization,” “appropriate style and tone,” “mechanical correctness.” Every one of these can be reliably produced by current AI. For quantitative reasoning: “interpret mathematical models,” “solve multi-step problems,” “explain mathematical concepts.” Again, AI delivers consistently. The only gen ed outcomes that AI arguably hasn’t mastered are those requiring physical presence (like public speaking) or genuine human interaction, precisely the kinds of collaborative, hands-on learning experiences that should now define university education.

The Current State and Its Vulnerabilities

Our environmental science student’s predicament reveals the fundamental disconnect in American higher education. State boards of education operate using decades-old assumptions in mandating that she complete a prescribed menu of gen ed courses. Access to information was limited. Expert instruction in a major would ensure competency but all graduates needed general knowledge to be well rounded citizens. The bureaucratic machinery supporting this system – assessment committees, learning outcome rubrics, transfer agreements – continues grinding forward, seemingly oblivious to the AI revolution that has transformed how students like her engage with general knowledge.

Consider how she has already demonstrated the standard five core gen ed requirements. Her application essay on Greenland ice cores showcased sophisticated analysis of scientific data, the kind of quantitative reasoning and critical thinking that general education courses are for. The AI system helped her find relevant research, structure complex arguments about climate change patterns, integrate statistical evidence, and connect historical data to current environmental concerns. Why require the first-year writing and quantitative reasoning courses?

The system’s vulnerability becomes even clearer when examining specific course requirements. In her mandatory “Critical Thinking and Writing” class, she will be asked to analyze arguments, identify logical flaws, and present counterarguments, all tasks that AI already guides her to perform with remarkable sophistication. Her required “Quantitative Methods” course will repeat systematic problem-solving approaches that AI systems already model with step-by-step clarity. Even in courses focused on civic and social responsibility, where she’s supposed to develop ethical reasoning skills, AI can generate nuanced discussions of environmental justice and craft compelling arguments about climate policy.

Perhaps only those of us who use AI regularly and know the gen ed ecosystem well see the problem here: that the current system’s approach to accessing the world’s knowledge – teaching students to find, evaluate, and use information effectively – has been fundamentally disrupted by AI capabilities. When our student researched those Greenland ice cores, AI didn’t just find sources; it created the illusion of mastering them, synthesizing complex climate science into coherent narratives without engaging with the fundamental question that should be at the heart of her education: how do we actually know what we know about climate change? The AI smoothly integrated data about atmospheric composition, glacial dynamics, and temperature variations without illuminating the human processes of discovery, verification, and scientific consensus-building that make this knowledge reliable.

The Fundamental Challenge: Knowledge Production and Verification

Put another way: our environmental science student’s essay was good enough to get into college but was it good enough to pass a college course? When she used AI to write about Greenland ice cores, she received a sophisticated analysis drawing from decades of climate science research. The AI seamlessly integrated information about ice sheet dynamics, atmospheric composition, and temperature proxies. But here’s what AI can’t do well: it can’t explain how scientists first discovered they could reconstruct past climates from ice cores. It can’t demonstrate how researchers validate their dating methods. It can’t show how the scientific community builds consensus around climate reconstruction techniques. It can’t easily mine from its training data the key debates along the way to what information it gives her.

So what kind of courses does the environmental student need now? Not general education but a focused environmental science or scientific methods course, to do a deep dive into her field, to read the original papers that tell her more than AI. And yet she won’t graduate if she skips her gen ed requirements and goes straight to the 300-level course.

Wasting her time is the real crisis in higher education. States continue mandating distribution requirements across disciplines, assuming that exposure to content and development of basic skills are what students need. But our student doesn’t need help accessing or synthesizing information. AI already does that brilliantly. What she needs is something more fundamental: understanding how we know what we know about Earth’s climate history in the first place.

Universities must restructure themselves around three crucial questions that AI cannot effectively address:

• What do we know?

• How do we know it?

• What remains uncertain or unknown?

These questions are the same whether a student is going to be studying science, history, philosophy, engineering, poetry, or economics.

These questions are the same even for a student who never used AI in high school. The environmental science student represents one level of preparedness but all students in an AI world, even those who have not yet used ChatGPT or Claude, should be taught by faculty who will always articulate what knowledge is upstream of AI and what is general knowledge.

These questions demand hands-on engagement with the processes of discovery and verification. Our student needs to understand how scientists over the years have understood time and eras, how they learned to extract and analyze ice cores, how they cross-validate their findings with other climate proxies, and how the scientific community evaluates new evidence and builds consensus. She needs to grapple with the uncertainties and gaps in our understanding of past climate changes, areas where knowledge is still actively being produced rather than simply transmitted. She needs to be upstream of AI.

This upstream position, where primary research and knowledge validation occur, is precisely where universities must focus their educational mission. While AI excels at synthesizing existing knowledge, it cannot participate in the fundamental human processes of discovery and verification that create new knowledge. Whether in environmental science, historical research, literary analysis, or social systems study, universities must teach students to work at this upstream level—where knowledge is produced, validated, and critically examined before it becomes part of the vast data pool that trains AI systems. And isn’t creating new knowledge what we want our student to do?

The Necessary Transformation

The implications for the gen ed bureaucracy extend far beyond simple reform. A complete reorientation toward new knowledge production rather than general knowledge consumption is needed. Every structure built to mandate, assess, and accredit general education must be dismantled and rebuilt around the university’s essential upstream role. State boards can no longer justify requiring students to complete 30-45 credits of general education courses when AI can instantly achieve their stated learning outcomes. More importantly, these requirements actively delay students like our environmental science student from learning what new knowledge might be in their fields.

The entire apparatus of general education assessment – the committees, the rubrics, the sample collection, the reports – is inefficient and counterproductive. The current system effectively positions all students equally as downstream consumers of knowledge rather than as individual upstream participants in knowledge creation.

Dismantling current gen ed requirement to align with a new upstream mission will be hard. First, state boards must put a sunset date on all gen ed credit requirements that measure AI-replicable outcomes, replacing them with requirements focused on knowledge production and validation. Second, accrediting bodies must revise their standards to emphasize students’ engagement with primary sources, raw data, and fundamental research methods. Third, institutions must transform their assessment bureaucracies to evaluate how effectively students engage in knowledge creation rather than knowledge consumption. Finally, state transfer and articulation agreements must be completely rewritten to focus on demonstrable engagement with the upstream processes of research, discovery, and validation that AI cannot replicate.

The Political and Financial Reality

For our environmental science student to experience a transformed, individualized education, states must overcome staggering political and bureaucratic challenges. A state legislature would need to publicly acknowledge that its entire approach to public higher education has been rendered obsolete by AI, a politically fraught position that challenges decades of established policy. This acknowledgment would force lawmakers to confront an uncomfortable reality: the gen ed requirements they’ve mandated and funded for years no longer serve their intended purpose.

The complexity of this process will extend far beyond a single legislative session. Restructuring would need to survive multiple changes in political leadership, intense scrutiny from stakeholders (including faculty and departments whose very existence depends on the gen ed mandate) who benefit from the current system, and inevitable resistance from those whose positions and authority derive from existing gen ed structures. The administrators and committee members who currently oversee our student’s gen ed requirements would need to participate in dismantling their own bureaucratic apparatus. History tells us how this will go.

The financial implications match the political complexity. Back of the envelope: a medium-sized state system would need to invest $8-10 million just in the planning and approval phase, funding policy research, legal analysis, stakeholder engagement, and system redesign over 2-3 years before implementing any actual changes. For context, this initial investment represents roughly the cost of providing traditional general education courses to 2,000 students like our environmental scientist. Implementation across 4-6 public universities would then require another $12-15 million for faculty training, curriculum development, new assessment systems, and knowledge verification technology, plus $2-3 million in annual recurring costs. AI can do the math. It’s well in the hundreds of millions across the country.

These estimates, however, don’t capture the hidden costs of an overhaul. Faculty must be retrained to teach upstream knowledge production rather than downstream content and skill delivery. Administrative systems must be reconfigured to track and assess different kinds of learning outcomes. Most significantly, universities must maintain their existing gen ed infrastructure while simultaneously building new structures, ensuring that current students can complete their requirements while the system transforms around them. For our environmental science student and her peers, this means slogging through the required courses because that’s what the state wants.

The New Focus: Epistemological Inquiry

Imagine our environmental science student in a transformed undergraduate education system where epistemological inquiry drives every course. Rather than simply learning about climate change, she would systematically explore how scientists discovered they could reconstruct past climates from ice cores. Her education would focus on what AI can’t do.

In this transformed approach, her laboratory work would involve examining actual ice core samples and replicating foundational experiments in climate science. When studying temperature proxies in ice, she wouldn’t just learn the current scientific consensus, she would trace how researchers first discovered these relationships, what challenges they encountered, and how they verified their findings. Her professors would guide her through examining original scientific papers, understanding how methodologies evolved, and recognizing how each piece of evidence was validated before becoming accepted knowledge.

The role of citations and footnotes would take on new significance in her education. Rather than treating them as mere formatting requirements, she would understand them as trust markers, concrete links to the human researchers who developed the techniques she’s learning. When reading about the Greenland ice sheet, she would trace citations back to the original expeditions, understanding how early findings led to refined methodologies and eventually to our current understanding of Earth’s climate history.

This approach reveals the limitations in current AI systems and makes the case for higher ed’s relevance. While AI can process citations as text patterns, it fundamentally cannot understand their meaning as links in knowledge networks. When an AI helps our student write about climate change, it may generate plausible-looking citations, but it cannot grasp the actual chain of discovery and validation that makes climate science reliable. This limitation becomes especially clear when she encounters cutting-edge research or areas of scientific uncertainty, precisely the domains where universities must prepare students to work.

In her chosen field of environmental science, there are vast archives of climate data that haven’t yet been fully analyzed, new methodologies being developed for ice core analysis, and emerging questions about climate system interactions that even the most sophisticated AI cannot yet address. By focusing on how we know what we know about climate change, her education prepares her not just to understand existing knowledge but to participate in creating new knowledge in these unexplored areas.

Looking Forward: The Path to Implementation

Consider how our environmental science student’s education could change under a restructured system. Instead of spending her first year checking off general education requirements that AI can simulate, she would immediately engage with the fundamental processes of climate science knowledge production. Her writing course would focus not on basic composition, which AI handles effectively, but on understanding how scientists document their methodologies, validate their findings, and build consensus through peer review. Her quantitative reasoning requirement would involve hands-on work with climate data, learning not just to analyze statistics but to understand how researchers establish reliable methods for reconstructing past climate conditions.

This transformation requires an unprecedented coalition. University leadership must acknowledge that AI has fundamentally altered the landscape of gen ed. Faculty senates must rebuild curricula around knowledge validation rather than content delivery. State boards of education must revise assessment frameworks to evaluate students’ ability to engage with primary research and understand knowledge production. Regional accreditors must update their standards to emphasize upstream learning, where students work with raw data, primary sources, and fundamental research methods.

Most crucially, state legislators must champion this complex transformation, understanding that the workforce needs graduates who can participate in knowledge creation, not just general knowledge consumption. For our environmental science student, this means developing the skills to contribute new insights to climate science, perhaps discovering previously unexamined relationships in ice core data or developing innovative methods for analysis. The goal isn’t to help her compete with AI in summarizing existing research but to prepare her to expand human knowledge in ways that AI cannot.

To implement these changes, states must completely restructure their approach to undergraduate education around epistemological skills and individual knowledge creation. This transformation isn’t about superficially reimagining education, it’s about fundamentally reorienting public universities toward their essential role upstream of AI systems, where each student learns to participate in creating and validating knowledge. Only then will public universities fulfill their mission in an AI-enabled world: developing graduates who can work upstream of AI, where knowledge originates through individual discovery, rigorous validation, and scholarly advancement. The general education system as we know it may not survive this transformation, but something more valuable will take its place: an educational model that positions each student to expand human knowledge rather than simply consume it.