Thread Careers
← All articles

Fighting AI With AI: Why the Outrage Over Online Degree Cheating Misses the Point

August 10, 2026 · 5 min read

A recent New York Times piece raised alarm about online college courses being quietly hollowed out by AI, students letting chatbots do the reading, the writing, and the thinking, while a credential gets printed at the end anyway. The framing is familiar by now: a crisis of academic integrity, a generation that never learned to think for itself, a degree that means less than it used to.

That framing is not wrong, exactly. But it is incomplete in a way that starts to feel almost willfully naive once you look at the full picture students are operating inside of.

The deal was already broken

College in the United States is not cheap, and it has not been getting cheaper. Total student loan debt in the country sits above 1.8 trillion dollars, and the average bachelor's degree holder leaves school owing somewhere in the neighborhood of 30,000 dollars, with graduate degree holders often carrying six figures. That debt follows people for a decade or more, shaping when they can buy a home, start a family, or save anything at all.

The traditional justification for taking on that debt was simple: a degree bought you access to a stable, well paying career. That promise is fraying. Recent research paints a fairly stark picture of the entry level job market right now. A Stanford Digital Economy Lab analysis found employment for workers in their early twenties in AI exposed roles has fallen sharply since 2022, with steeper declines in fields like software development. A separate Harvard study of tens of millions of workers found junior employment dropping at companies that adopted generative AI while senior employment held steady or grew. Entry level job postings overall are down significantly since 2023, and underemployment among recent graduates, meaning graduates working jobs that never required a degree in the first place, is at its highest point since the pandemic.

So the arrangement many students are working within already looks like this: pay tens of thousands of dollars, take on years of debt, in exchange for a credential that increasingly does not guarantee the job it was supposed to unlock, in a market where the first rung of the career ladder is the one AI is eating first.

The double standard

Meanwhile, once those same students land a job, if they land one, a growing number of employers are not just permitting AI use, they are mandating it and measuring it. Meta has made "AI driven impact" a formal, scored part of every employee's performance review. Microsoft has told managers that AI use is no longer optional. Google's leadership has delivered the same message internally. Amazon tracks individual AI tool usage through an internal system and factors it into evaluations. A 2025 survey found more than half of companies now require a certain number of employees to actively use AI tools, and some, like Accenture, have reportedly tied AI usage to promotion eligibility.

So the same institutions, and the same economy, telling eighteen year olds that using AI to complete their coursework is dishonest are, three or four years later, going to tell those same people that not using AI aggressively enough is a performance problem. That is a genuinely strange thing to ask someone to hold in their head at the same time.

Where the "fighting fire with fire" logic actually breaks down

Here is the honest complication, and it is worth sitting with instead of waving away. A college degree is not currently structured as a credential that certifies AI proficiency. It is structured, at least on paper, as a credential that certifies a person can think, reason, write, and problem solve without assistance, and that those underlying capacities transfer to situations no AI tool has been trained for. If a student outsources that entire process to a chatbot, they may still get the piece of paper, but they have not built the judgment the piece of paper claims to represent. That gap does not disappear. It just gets deferred to the first moment at work when a genuinely novel problem shows up and there is no AI shortcut sitting there ready to solve it.

Several education writers covering this shift have made a related point: some of the courses students are most aggressively using AI to skip are ones built around rote, formulaic assignments in the first place, work that was already testing compliance more than comprehension. That is a real critique of how some curricula are designed. It is not the same as an argument that foundational thinking skills stopped mattering.

What this actually points to

None of this means students letting AI do their coursework are making some savvy, forward looking career move. Many of them are just avoiding work, and the honest ones in that Times piece and elsewhere seem to know it. But the anger directed almost entirely at students, while the institutions charging six figures for a degree and the companies now mandating AI fluency go largely unexamined, is aimed at the wrong target.

The more useful question is not how to catch more cheaters. It is why we are still asking twenty year olds to pay a small fortune for a credential built around a set of skills that the job market on the other side of graduation is simultaneously automating and, in the same breath, telling them to lean into harder. Something in that pipeline, the cost of the degree, what it is meant to certify, and how honestly it is now assessed, needs to be redesigned around the world as it actually is, not the world college pricing still pretends we live in.

Sources