The United States, long regarded as the world’s dominant power, has long prided itself on technological innovations and universities that are a symbol of academic excellence. With 420 Nobel laureates, it leads all nations in total Nobel Prize recipients. However, for the first time, the academic achievement levels of students in the US have decreased. Instead of consistently increasing our literacy and comprehension rates (as we have steadily throughout our country’s history), the youngest generations have regressed from their predecessors in terms of the amount of knowledge and problem-solving skills they have. Experts have blamed the pandemic for the lapse. In reality, the cognitive decline of US students has been noticeable since 2009.

The University of San Diego has frequently ranked as one of the top public universities nationwide. A recent report from the UCSD Academic Senate, however, documented a staggering increase in incoming freshmen who required remedial math education. Between 2020 and 2025, the number of below-age level freshmen increased from less than 1% to about 12.5%. The university uses a split remedial math course system, offering non-credit courses that revisit elementary and middle school math before students are deemed ready for college mathematics. This means that at top universities, like UCSD—ranked No.#29 in National Universities—students struggled with middle-school math. Notably, of the students placed in this remedial course, 25% had a 4.0 high school GPA in math, meaning that they got all As in high school-level math, and 20% had already taken a Calculus course in high school.

In what Rose Horowitch, a journalist at The Atlantic, calls a “recipe for idiocracy,” the efficiency of K-12 education has noticeably decreased since the pandemic. The Nation’s Report Card documented that math scores among 13-year-olds declined by 14 points between 2020 and 2023 compared with levels a decade prior. Horowitch and educators pointed to the “attention-shredding” nature of smartphones and students’ reliance on AI large language models and calculators as the drivers of this drop. Yet, graduating classes have had their average Math (from 3.3 in 2010 to 3.41 in 2022) and English (from 3.19 to 3.36) GPAs consistently driven upwards each year.

The A.I. Problem

A 2024 study published by the National Academy of Sciences found that students using AI “tutors” like GPT-4 initially performed 127% better, but their performance plummeted by 17% once the AI tools were removed from their learning, indicating that unsupervised access to AI causes detriments to learning. Similarly, research published on IFLScience found a negative correlation between AI usage and critical thinking, suggesting that by allowing students to bypass the “productive struggles” which shape our students into more productive thinkers, AI is directly harming students’ capacity to think. Large language models are outsourcing the process of thinking itself, in what researcher Evan F. Risko termed cognitive offloading, something undoubtedly dangerous to the professionals and great thinkers of the future because with only one prompt AI can replace their jobs and our diversity of mindsets.

Some in Silicon Valley say they want to see their tools used responsibly and in ways that maintain integrity in education, as OpenAI claims it is building “tools for educators” that should be helping them facilitate their students’ learning. However, a year ago it developed a technology that was 99.9% accurate in detecting work generated by ChatGPT. The project had been stuck in internal debate at the company for roughly 2 years. OpenAI would decide against making it public, knowing that people would stop using ChatGPT and switch to a competing product without a verified AI detection method.

Grade Inflation

All of these new tools, many marketed as “for learning”—including generative AI tools, calculators, smart phones—have burdened long-term student success rates in tandem with another factor: grade inflation. Apart from its general connotation, grade inflation takes two forms (both coined in a study by University of Texas associate professor Jeffrey Denning): mean grade inflation and passing grade inflation. Both, Denning argues, have had a detrimental toll on student success into adulthood via lower adult income, post-secondary enrollment rates, and even graduation rates (counterintuitively reversing the inflation’s intended effect). Overwhelmingly, grade inflation allows mediocre students to do less work, while making it harder for hard-working students to distinguish themselves academically.

Passing grade inflation is the phenomenon of more students receiving higher average grades for the same or lower quality of work. This has failed to bring students to succeed in their post-secondary tracks; more students enroll in postsecondary schools (especially two-year colleges), but they don’t all graduate. On the flipside, mean grade inflation is the average rise of grades (typically measured across a graduating class), and it has even more of an impact on the higher education level, where more and more students (75%)  are graduating without the proper skills to perform well in their occupations. However,  teachers and professors are starting to use other grading methods, such as behaviour or engagement during class, rather than test-based evaluation.

According to a report from Harvard’s Office of Undergraduate Education, one factor responsible for grade inflation is that faculty members worry that tougher grading could lead to unfavorable student evaluations, which might hurt their performance reviews and future job prospects. On that note, a professor also explained that some of the pressure to give better grades comes from other faculty members. Since students tend to pick classes where they know they will succeed, professors lower their standards to raise enrollment as they compete for students. And this has happened in colleges across the United States, as half of the respondents from a survey had not taken a single liberal arts course that required writing more than 20 pages total. At the same time, the average number of hours a student spent studying declined by almost 50%, from 25 to 13 hours per week, and the average GPA of Harvard’s 2024 class was 3.8, which is a 0.75 increase compared to 50 years ago.

In response to that, on February 6, Harvard’s Amanda Claubaugh, Dean of Undergraduate Education, emailed students and faculty with a proposal for reforming the grading system. The 19-page document includes a set of recommendations intended to “restore grades to their role as meaningful indicators of student performance and feedback.” The main recommendation was that instructors limit the number of A grades to 20% of the students enrolled in a class, plus four additional A grades. The report stated that those four additional A grades would have a bigger impact on classes that have “an unreasonably low number” of students, like a seminar with 15 people, for example. Another change is that they would stop using GPA and instead use the average percentile rank (APR). APR shows the percentage of students at or below a particular student in the class—measuring students’ academic performance relative to each other. This calculator works in the same way as percentage ranking in the SAT (someone with a 99% scored better than 99 out of 100 test takers). These changes are not yet official policy. Right now, the proposal is under deliberation, and the administration is soliciting feedback from students and faculty. If advanced to the next step, it would become a legislation, which would be voted on by the full faculty of Arts and Sciences.

The problem isn’t just teachers: part of it can be attributed to the fact that grade inflation has increasingly become the norm in educational institutions across the US. Students, parents, and teachers are all expecting higher grades to come out of less and less work. The struggle to validate our educational system as successful has transformed from a measure of how much kids are learning to how many kids we can say learned. Metrics of success have become more and more impersonal; increasingly so as districts compete for students for funding and students compete for collegiate admissions. 

The incentive for students to succeed for their own futures has been perverted by institutions’ clamoring for government funding. 

At the same time, the UCSD case study presents another outlook: students from under-resourced schools. LCFF+, a classification for California institutions with high concentrations of low-income and foster youth, made up 53% of those in remedial math, despite being only 20% of the incoming class. After their 2021 transition towards a “test blind” admission policy, the school saw a significant rise in students from LCFF+ schools: in 2024, 25% of the incoming class and 31% of in-state students were admitted from under-resourced schools, compared to 13% and 17% in 2020, respectively. 

So grade inflation, as it’s marketed, may in fact be successful in pulling up lower-income students who have trouble accessing the same resources students from fully-funded schools can count on. It may be diversifying the collegiate population, which is likely beneficial in making the workforce more well-rounded and invested in the future of all populations. But—and an important hesitancy on the part of many analysts—as these populations come into college, into the workforce, there is no good they will do without the ability to think. If the people we rely on to be our politicians, our lawyers, our doctors, our essential workers, and the people that service every aspect of our lives cannot make decisions without consulting an outside entity, we have failed to prepare them for society and for any future with or without technology. 

The solution doesn’t reside in reverting to the old ways of grading, where vindictiveness and singular test scores determine a student’s future. It also can’t be found in our current system, or even a more exaggerated version of it, where every student is marketed as successful. Educators, and the education system itself, have always relied on compromise. Finding a happy medium—where students are encouraged to envelop themselves in the learning process while also creating successful products—will be the key to a successful society. 

Looking Forward

At some point, though, a shift must happen. It won’t be quick—we have a lot to catch up on—but it will occur. And we have to start now. As we catch UCSD students up on their remedial math skills, we must also ensure our students in elementary are learning math. Making the whole system shift, rather than doing it incrementally, means greater proportions of our students will learn to be successful and—more importantly—to think, especially in a society where our actions are outsourced and critical thinking becomes a rarity.