The days of using AI to cheat on your exams might be officially over.

@chuma_the_ai_coach

7.7K followers · usually 2.2K views a post

Views

29.2K

Likes

906

Comments

41

Shares

57

Why it worked

The video presents a compelling real-world example of AI's impact on education, using a dramatic score drop to highlight the issue and sparking discussion about academic integrity.

Summary

A Brown University economics professor noticed unusually high scores on a take-home midterm exam, suspecting students were using AI to cheat. When he moved the final exam offline, the average score plummeted from 96% to 48%, revealing the extent of AI-assisted cheating.

Structure

1Professor suspects AI cheating due to high midterm scores.
2Professor moves final exam offline.
3Final exam scores drop dramatically.
4Analysis shows a stark difference between online and offline performance.
5Discussion on the implications for AI in education.

Product placement

ChatGPT (AI tool that generates text, used by students to cheat on tests, merely appears as a logo in the video, removing it would not change what happens in the video) Claude (AI tool that generates text, used by students to cheat on tests, merely appears as a logo in the video, removing it would not change what happens in the video)

On-screen text

TAF A PROFESSOR SUSPECTED STUDENTS WERE USING AI TO CHEAT ON TESTS, SO HE MOVED THE EXAM OFFLINE AND SCORES DROPPED FROM 96 TO 48. TAF

Original caption

No more cheating on exams for these students 😂😂😂😂 #ai #chatgpt #aicheating #claude #openai

More from @chuma_the_ai_coach

Their other posts in the index, biggest breakout first.

Transcript

The days of using AI to cheat on your exams might be officially over. Have a look at the story from Professor Roberto Serrano who noticed something strange with the results of students where the exam average reached 96% far above the course's usual 65 to 80% range. After noticing similar and AI-like answers, Serrano moved the final exam offline. The graph shows Brown students' take-home midterm scores clustered near 100, while in-person final scores dropped sharply. After Serrano made the midterm exam, 10 students chose the class and nine students did not take the exam. The scores of the remaining 29 students are displayed here. So the good professor said, no, something is up here. So what they did was that they took the exam offline and required students to write without any computer assistance. And the results were remarkably different. The average mark went from 96% all the way down to 48%. A completely different picture. I think this is a cause for concern and maybe a different or new way that lecturers might consider to clamp down on the over-reliance on artificial intelligence. Or we are facing a situation where we need to reconsider how we need to test for knowledge. As I'm one of those people who are not against the use of AI in education.