The era of large models is forcing a transformation in human education, and it also compels us to reflect on what sets us apart from AI.
Wrongly Caught in AI’s Crossfire: How Do I Prove I Didn’t Cheat With AI
A couple of days ago I came across an article: https://mp.weixin.qq.com/s/GQdnxdyJ-L4pijcuYAf5rA
The article mainly discusses the conflicts between students and teachers over grading assignments in the AI era, and how to clear your name after being wrongly flagged for cheating by AI. Here are the details:
- Students wrongly flagged for cheating by AI detection tools
- A student at the University of Houston, Leigh Burrell, carefully completed an English assignment that her professor suspected was written with AI and gave her a zero. It took her more than half a month of submitting 15 pages of writing-process screenshots, timestamps, and video of her doing the assignment to clear her name. Afterward, to avoid being flagged again, she set up a camera to record the entire process whenever she did her homework.
- A public health student at the University at Buffalo independently completed an assignment that Turnitin’s AI detection tool wrongly flagged as “possibly AI-generated.” The school launched an academic misconduct investigation based solely on that, and such cases are hard to appeal — the burden of proof falls entirely on the student. Groups like non-native English speakers and students with disabilities are more likely to run into this.
- Some students were wrongly flagged for using AI twice in the same course. Another student poured their heart into a meta-analysis paper, and 60% of it was wrongly flagged by Turnitin as AI-generated — and they had no idea how to rewrite the flagged content in their own words.
- One student used Visual Studio Code to complete a web development graduation project. Despite having all kinds of records documenting the development process, the professor doubted everything and refused to accept that they hadn’t used AI.
- Different extreme cases of AI use
- Chungin “Roy” Lee, a computer science student at Columbia University, used AI to “reference” almost every liberal arts assignment, and also built an interview-assistance tool for programmers and launched a successful startup.
- One teacher used AI to prepare lessons, and instructions meant for ChatGPT showed up in the lecture notes. Students complained — yet that teacher’s syllabus prohibited academic misconduct such as using AI.
- Problems with AI detection tools: Many AI detection tools today check “whether something was generated by AI,” but their overall accuracy is low, which is driving teachers and students further and further apart.
- The frustrations and conflicts on all sides
- On the student side: “good students” get suspected because AI detection tools flag them or because their writing style is “too perfect,” and clearing their name is hard. Some students use AI reasonably to assist their learning, while others rely on it excessively or even cheat.
- On the teacher side: faced with a flood of assignments that “sound like AI,” it’s hard to tell what’s real. Some teachers use AI to assist their teaching only to be called hypocritical by students, and different teachers draw the line at different places when it comes to how much AI students may use, which easily sparks conflict.
- Current workarounds
- For students: for coding assignments, keep a rich git commit history (typos, failed builds, and all); for written assignments, make good use of the version history in Word and Google Docs, or record a video or livestream of yourself doing the work — always be mindful of preserving the “evidence” of your writing process.
- For teachers: some believe the only way to ensure fairness in written assignments is to go “old school” — handwritten, in person, with a human proctor, such as writing essays in class by hand under the supervision of a real instructor.
Reshaping Originality and Human Uniqueness
What’s most thought-provoking for us is this: should the way we educate evolve along with technology? If AI can do these things well, why assign work that AI can do well? Where does human uniqueness lie — and shouldn’t assignments incorporate that human uniqueness?
The era of large models is forcing a transformation in human education, and it also compels us to reflect on what sets us apart from AI.
I think, first of all, assignment design needs to change. Stop with the template-driven, mechanical assignments — let students exercise their creativity, critical thinking, and emotional experience. Take essay writing: stop making students write the same cookie-cutter prompts like “My Dream” or “An Unforgettable Day.” Those prompts have been written to death. Let students draw on their own life experiences and write something with more depth and personality, like “How I Grew From a Failure” or “My Unique Take on a Social Phenomenon” — prompts that allow for individual differences.
It’s like when I interview people: I tend to focus on asking how they solved the bigger challenges they ran into in previous projects.
In writing instruction, students no longer just need to master grammar and rhetorical technique — they need to develop critical thinking, creativity, and the ability to express emotion. These are things AI still can’t replace.
Second, the education assessment system needs to be rebuilt. It can’t just look at scores and outcomes anymore — it has to pay attention to students’ thinking process and growth process. It’s like building a product: you can’t just look at whether it sells well in the end; you also need to look at what problems came up during R&D and how they were solved. You can use portfolio assessment, collecting students’ drafts, revision records, and reflection journals, and evaluate their abilities holistically.
Beyond that, teachers need to transform and upgrade too. They have to go from traditional purveyors of knowledge to coaches and guides for students’ growth. They need to learn to use AI tools well and integrate them into teaching rather than resist them. It’s like running a business: you can’t reject new technology — you have to embrace it and make it work for you.
AI is not the enemy of education — it’s an opportunity to rebuild it. We have to seize this opportunity to make education more humane and better at cultivating talent that is distinctive and creative. I hope that in the future, education will let every student find their own value and stand firm amid the AI wave rather than being drowned by it.
At the same time, humans must re-examine their own uniqueness — we are not just receivers and processors of information, but also bearers of emotion, judges of morality, and transmitters of culture. In an era where AI can generate articles, compose music, and even do artistic design, the unique value of humans lies in our perception of the world, our understanding of life, and our thinking about ethics. Therefore, educational transformation is not merely the result of technological drivers — it’s an inevitable choice for human self-understanding and future development.