Title: Students Get Higher Grades with AI, But Learn Less

Which is more important, learning or good grades? You might think that one is the same as the other, but a study has found that student AI use could be resulting in higher grades but less learning. A survey by Times Higher Education found that around 85% of university students use AI for their coursework in some way, with 25% using it to complete assignments and 19% using it to write full essays.

To better understand AI’s impact on learning, researcher Igor Chirikov looked at syllabi and grades from 319 courses across 84 departments at a university in Texas. His data covered 2018 to 2025, so it included years before and after AI became easily accessible. Chirikov identified the university courses that were most likely to be affected by AI as those with high levels of unsupervised tasks, like writing and computer programming. Courses with high exposure to AI have seen A grades — the top level — rise in recent years. In 2022, the year ChatGPT was released, the average share of A grades in one of these courses was 44%. By 2025, this had risen to 48%. At the same time, the number of lower grades fell. The average share of C grades went from around 8% to 7% between 2022 and 2025, while lowest-level D and F grades went from 3% to 2%.

For comparison, the study also looked at grades for oral presentations, where AI isn’t very useful. Here there were no grade changes, supporting the idea that AI is the reason for improvements in classes with more unsupervised tasks.

There are two problems with these AI-inflated grades. First, they may not be the result of real learning, and students could be graduating with weaker skills. Second, while good grades often help graduates get hired, AI inflation could reduce their value to possible employers — because if everyone is getting high grades, what do those grades really mean?

📖 Understanding the Article（3楼）

These questions check your understanding of the article’s main points.

1. In your own words, what is the main finding of Chirikov’s study? What is the evidence that AI, rather than something else, is causing the grade changes?

2. The article uses oral presentations as a comparison. Why was this a clever choice? What does it prove?

3. What does the article mean by “AI-inflated grades”? According to the author, what are the two problems with them?

4. The study covered 2018 to 2025. Why was it important to include data from years both before and after AI became widely available?

💬 Your Experience（4楼）

These questions invite you to share your own experiences and observations.

5. The survey found 85% of university students use AI for their coursework. Does this number surprise you? From what you’ve seen among classmates or friends, what would you guess the real rate is?

6. Have you ever used AI tools (ChatGPT, translation apps, etc.) to help with schoolwork? If yes, how did you use them? If no, why not?

7. Where do you personally draw the line between “acceptable AI help” and “cheating”? Can you give a specific example?

8. Think of a subject where your grades felt like an accurate reflection of your real ability, and another where they didn’t. What made the difference?

9. If a classmate got an A using AI while you got a B through honest work, how would you feel? What would you actually do?

📝 Words & Language（5楼）

These questions focus on the language used in the article.

10. The article invents the phrase “AI-inflated grades.” Why is “inflated” an effective word here? What image does it create? Can you think of something else in daily life that gets “inflated”?

11. The word “unsupervised” appears in the phrase “unsupervised tasks.” What does “unsupervised” suggest to you? What kind of school tasks are supervised, and what makes the difference important in this study?

🎯 Debate & Beyond（6楼）

These questions encourage you to take a position and defend it.

12. If AI helps students get higher grades but learn less, who is most responsible — students who use it, teachers who don’t prevent it, or AI companies that make the tools?

13. The article treats AI-inflated grades as a problem. But some people might argue: if grades were already an imperfect measure of real ability, maybe AI is just exposing an old problem faster rather than creating a new one. Which view do you lean toward?

14. Should schools try to ban AI use in coursework, or should they change the way they assess students instead? What would a better solution look like?

15. The share of A grades rose from 44% to 48% between 2022 and 2025. Do you think this trend will continue in the next few years, level off, or reverse? Why?

16. If you were a teacher designing an assignment that AI cannot easily complete, what would it look like?

📚 Vocabulary List

A. Academic & Research Terms

1. coursework (n.) — Work done by students as part of a course of study.

“…around 85% of university students use AI for their coursework…”

2. syllabus (n., plural: syllabi) — An outline of the topics covered in a course.

“…looked at syllabi and grades from 319 courses…”

3. unsupervised (adj.) — Not watched, checked, or monitored by someone in charge.

“…courses with high levels of unsupervised tasks, like writing and computer programming.”

4. accessible (adj.) — Able to be reached, entered, or used.

“…before and after AI became easily accessible.”

5. exposure (n.) — Contact with something, especially over a period of time.

“Courses with high exposure to AI have seen A grades rise…”

6. comparison (n.) — The act of examining two or more things to find similarities and differences.

“For comparison, the study also looked at grades for oral presentations…”

7. assignment (n.) — A task or piece of work given to students as part of their studies.

“…25% using it to complete assignments and 19% using it to write full essays.”

8. oral presentation (n.) — A spoken talk given directly to an audience, without relying heavily on written text.

“…grades for oral presentations, where AI isn’t very useful.”

9. graduate (v. / n.) — (v.) To complete a degree or course of study; (n.) A person who has completed a degree.

“…students could be graduating with weaker skills.” / “…good grades often help graduates get hired.”

10. employer (n.) — A person or organization that pays people to work for them.

“…reduce their value to possible employers…”

11. AI-inflated (adj.) — Made to appear higher or better by AI, without reflecting real improvement. (A term created by the article.)

“There are two problems with these AI-inflated grades.”

12. inflation (n.) — A general increase in something that reduces its real value. (Usually used in economics; here applied to grades.)

“…AI inflation could reduce their value to possible employers…”

13. identify (v.) — To recognize or determine exactly what something is.

“Chirikov identified the university courses that were most likely to be affected by AI…”

14. result in (phrasal v.) — To cause or lead to a particular outcome.

“…student AI use could be resulting in higher grades but less learning.”

15. in some way — To some extent; not completely or fully.

“…85% of university students use AI for their coursework in some way…”

16. for comparison — Used to introduce information that serves as a basis for comparing.

“For comparison, the study also looked at grades for oral presentations…”

Try to use some of these words in your discussion posts!

1. In your own words, what is the main finding of Chirikov’s study? What is the evidence that AI, rather than something else, is causing the grade changes?

The main finding of Chirikov’s study is students will have better grades with the help of AI assistant. The evidence for that is mor students can get A grade after the release of ChatGPT

It doesn't surprise me. Although I'm not a college student, I've noticed many students younger, like students in senior high schools, use AI to assist their academic work. In my opinion the real rate could be even higher.

6. Have you ever used AI tools (ChatGPT, translation apps, etc.) to help with schoolwork? If yes, how did you use them? If no, why not?

I've used them and found out they are very helpful. For example, I turn to Deepseek for difficult problems or article polishing, and it can give a clearer answer than most people can.

1. In your own words, what is the main finding of Chirikov’s study? What is the evidence that AI, rather than something else, is causing the grade changes?

Main finding: depending on if AI is useful to the mission, the AI's impact on learning varys.

Evidence: based on a contrast analysis between students using AI and students not using AI to finish their courses. This analysis shows that when the tasks can be easily finished by AI, students' grades go up. However,as for those tasks that AI cannot easily do, for, there was no change of the grades.

Because AI cannot replace the students to make an oral presentation, though it can help students write drafts.

It proves that it is AI that improves students' performances in unsupervised tasks more.

When using AI, students can get better grades than those who don't use AI, because AI can outupt things above what the average do.