As Students "Flex" AI Use, Professors Worry They're "Outsourcing Thinking"

Are students using AI for homework, assignments etc isn't really the question anymore, Dr Prerana Srimaal, Head of Liberal Arts Department at Christ University, Bengaluru tells NDTV. "Of course they are. It would be naive to think otherwise. The question isn't if, it's how and to what extent."

For years, universities tried to catch students cheating. Professors and educators say the challenge is different now. AI has become so deeply embedded in student life that the question has gone beyond whether assignments contain AI assistance. It now revolves more around where institutions draw the line between ethical, legitimate help and academic dishonesty.

Srimaal says AI use is widespread among students but not always in the way people assume. "It's not just copy-paste cheating; it's also subtle dependence: structuring answers, generating ideas, polishing language."

Dr Sanchita Khurana, Assistant Professor at Mata Sundri College for Women, Delhi University (DU) tells NDTV that everybody does it now. "The worrying part is, it's not something worth thinking twice about now. There is a brazenness around taking 'help' from AI."

Since "everybody" has hopped on the bandwagon, students believe they will be at a disadvantage if they don't use AI. "Completely distancing yourself from AI is a huge disadvantage because when you use it for organising your ideas and making decisions, you save time, money, and effort. Whether for assignments or day-to-day tasks, students who do not use AI are somehow still at a disadvantage," Lakshita Gajendra Babu, a student of Christ University, Bengaluru says.

So students openly talk about using AI now? The answer seems to be an overwhelming yes. 
"We are very open about the usage of AI. It's become a flex of some kind to say that I used AI to solve this problem of mine or that I used AI to give me daily validation. Earlier people were more secretive about it, but now almost everyone uses AI in some form, whether for notes, coding, presentations, emails or studying. It's become pretty normal in college conversations," says Sonika Krishnan, another student from Christ.

Even those who aren't the biggest fans of AI end up using it once in a while. "I personally try to avoid using it (AI) as much as possible, so maybe once in a while," says Bhuvi Sharma, a student of English at Delhi University, who uses AI for " last minute revisions."

If AI Is Everywhere, What Counts As Cheating?

With AI now being omnipresent which is well established there remains a very thin line between assistance and cheating. "If AI is helping you think, it's assistance. If it's thinking for you, it's cheating," Srimaal believes.

Students admit the line can get blurry sometimes while adding that a 'copy-paste' job is clearly across the red line. "I personally draw the line when someone simply copies and pastes the entire assignment question into ChatGPT and then pastes the response back into a document and submits it without even proofreading the answers. Helping is when you put in your personal effort and then use AI for assisting your document and doing the basic needful," says Babu.

Then there are students who believe any kind of content creation using AI lies beyond ethical lines. "When AI is used to create texts from scratch, such as to write a story based on a mere prompt given, I would consider that unethical, and by extension, cheating. I have always seen AI as an assistant that must only be asked to streamline one's tasks  based on work that is predominantly carried out by the human, says Pranati R Narain, a student of Christ.

Professors admit that the cheating debate is rather subjective as they don't have a universally accepted definition around AI yet. DU professor Khurana raises an important point here that while academia navigates the AI wave, the tech may help bring "linguistically weaker" students at par with those who have been more exposed to the English language.

"As a teacher of English language, I am still trying to understand the implications of AI with regards to a lot of questions around language originality versus assisted language. There is also something to be said about rethinking what will constitute creativity in art now. 

"I cannot say for sure what that line between assistance and cheating is but going by my old school use of sources for writing a paper, perhaps one could use AI for gathering information and research, rather than relying on it for generating your opinions, which I think is intellectual cheating, at least," says Khurana.

Researching on AI platforms can however, sometimes be tricky as it is prone to hallucinations and can throw up entirely made-up material and pass it on very convincingly as facts. So it's imperative that one cross-check numbers, facts, case studies while researching on these platforms.

The Tell-Tale Signs

Professors say there are tell-tale signs even before they put a student's paper through software checks such as Pangram Labs.

"Certain catchwords which are popular with ChatGPT, for instance, like.. tapestry, quietly. There are certain ways of phrasing, especially the 'not this, but that' format. In general, an easy use of the superlative expression, suited very well, if I may add, to the hyper-positive focus of social media. The obvious em dash too. 

"It is also clear to us a student has used AI when the language is not theirs, I mean I know how an 18-21 year old of today's time can write; how you know if it's AI and not plain old plagiarism is that the language will be refined but not substantial; there will be free flow of verbosity but nothing important is really being said," Khurana points out.

However, students seem to have found workarounds to some of these clear tells. "I don't use the em dash anymore," says Bhuvi Sharma. 

Also because this is largely an AI-native generation, they are really well-versed with how these AI platforms behave. "As undergraduates, we can visibly identify which AI tool someone has used, whether it is Claude, Perplexity, Gemini, or ChatGPT. We have gotten really good at identifying it, and talking about it is not a secret anymore," says Babu. Like Krishnan earlier, Babu too says it's a big "flex" among students. Beyond academics, Babu points out that many students use AI for validation purposes and emotional advice as well.

Humanising The Machine

So do students try to 'humanize' AI-written work? "Definitely," says Krishnan. "People rewrite sentences, add small mistakes, simplify vocabulary or paraphrase content to make it sound more natural. It's kind of funny because now students spend time trying to make work sound less perfect," she says.

Narain, shares an interesting term - "de-AI" in this context. She says that "students very often attempt to humanize or 'de-AI' work that is written by AI models, mainly through re-writing sentences or through paraphrasing tools like Grammarly or ZeroGPT."

While teachers pick up smarts to catch students in the act, the latter may be picking up workarounds at an even faster pace turning the whole thing into a never-ending cat-and- mouse game.

Homework Will Never Be The Same Again

So has AI changed the way professors now design assignments? Completely, they say. 
"You can't design assignments today as if AI doesn't exist. We've moved towards process-based work, reflections, field engagement, and oral components," says Srimaal. Khurana of DU says AI has indeed "absolutely" changed the way assignments are handed out. "I have completely shifted to class testing on spot and without a phone in tow for most of my courses," she says.

Aakash Choudhry, co-founder of the Aakash Educational services (famous for Aakash Institute) and Founder and MD, Sparkl Edventure is a veteran of the coaching and education industry. He believes that homework can no longer remain a mechanical exercise of copying answers or solving repetitive questions without understanding. "AI has exposed the weakness of that model." The purpose of homework must now shift from answer generation to learning verification, he says while sharing a very interesting example of how to do that. 

"Instead of asking a student to simply write an essay, we can ask them to use AI to create a first draft, then critique it, improve it, identify weak arguments and explain their own position. In mathematics or science, homework can focus on process, reasoning, error analysis and application rather than just the final answer, he shares. The proof of learning matters more than completion of homework, Choudhry highlights.

The Deeper Fear: Are Students Outsourcing Thinking?

AI educator Ansh Mehra believes, "Over-dependence on AI can lead to intellectual obesity in our next generation, where they stop working on their thinking because they've offloaded their cognition to an LLM (large language models). He says till Class 5 or 6 AI use needs to be strictly restricted, while allowing some kid-friendly tools that help image or video generation, but not reasoning LLMs.

"Before AI, lazy students used to copy from their friends. The desire to skip effort is not new. It's just that AI has helped students fulfil that desire faster," Mehra warns.

Srimaal says while stronger students use it as a tool, weaker students tend to outsource their thinking to it, pointing to a worrying trend. Khurana paints a larger picture here saying what's playing out here with this dangerous learning trend goes beyond just AI. "This is not just due to the arrival of AI; it's also due to truncated attention spans, a superficiality of thought owing to the focus on appearances on social media," she points out.

"They are not learning anymore; there is a lack of engagement with their own thoughts, there is a lack of engagement with narrativity, hence no reading," Khurana adds.

On the other side, the students believe that AI is helping them learn better and also finish work faster. "It helps me learn better because, from experience, being asked to study 15 PDFs with 50 pages each, even with enough time, consumes a lot of time and energy. Doing it manually also includes reading and understanding everything, while AI does the first half of the work," Babu points out. Krishnan agrees and says, "It (AI) definitely helps finish work faster, especially repetitive tasks, but it also helps me learn better because it can explain difficult concepts in simpler and more conversational ways than textbooks sometimes do."

The Skills AI Can't Replace

Both students and educators seem to be on the same page about one thing - not using AI now feels like refusing to use the internet or Google twenty years ago. There is a clear consensus among students that the ones not using AI are at a severe "disadvantage". 

Choudhry points out that the student of the future will not be the one who knows the most facts, it will be the one who can combine human judgment with AI capability. "Therefore, education must move from 'What do you know?' to 'How do you think, apply, create and decide?'" he shares.

So what skills still matter? There is clear alignment here among academia.The keywords that kept popping during extensive conversations were - reading, writing, learning, speaking, critical thinking, prompt literacy, intellectual honesty, and problem solving. The student community believes that creativity, judgment, and asking better questions is going to set young people apart from the crowd as they enter the workforce.

As universities redesign assignments, and students learn how to "de-AI" their work, and while educators debate where assistance ends and cheating begins, one thing is fairly evident: AI is not leaving the classroom.

The challenge now is to ensure that while machines help students find answers, they do not stop them from learning how to think.