German tech influencer Lara Sophie Bothur shares student ChatGPT use cases from real conversations
Lara Sophie Bothur, German tech influencer and LinkedIn Top Voice in AI, shares insights on how students are using OpenAI’s ChatGPT for exam prep, structured learning, and critical thinking in education. Photo credit: Lara Sophie Bothur
German tech influencer Lara Sophie Bothur took to LinkedIn to share examples of how students are using OpenAI’s ChatGPT in academic settings, based on 50 anonymized conversations. The dataset, created in collaboration with the Q-Summit at the University of Mannheim, offers a current view of how AI tools are being used across higher education.
The examples show students using ChatGPT to break down complex topics, prepare for exams, and test their understanding. Rather than relying on AI to complete tasks, many prompts focus on guidance, structure, and feedback, pointing to a shift in how students are integrating AI into their learning workflows.
Students use AI to guide thinking and problem-solving
Across the shared conversations, students repeatedly ask ChatGPT to support their thinking without removing the challenge of the task itself.
Prompts include:
“I’m completely stuck on this task... can you explain how I should approach it without just giving me the answer?”
“You are my professor... give me a realistic dilemma... don’t give me the solution... challenge me with critical questions...”
“Ask me an exam questions and I’ll answer them.”
“Turn this into a cheat sheet I can revise in 5 minutes.”
Students also request structured academic support for research and writing:
“Formulate 3 very different research questions... create an outline (5 chapters) for the most interesting one... and name 3 potential pitfalls...”
“Please provide three arguments, including supporting evidence... presented in the style of an argumentative essay.”
There are multiple examples of students asking for clarity and adaptability in explanations:
“Explain this concept as if I were 12 years old and then again at university level.”
“Summarize this so I can quickly understand it again for my exam. Structure it clearly and highlight the key ideas.”
“Can you present this as a table or a mind map?”
The dataset also shows students using AI to test and refine their own work:
“Please show me exactly where there are spelling, grammar or logic mistakes. Don’t rewrite the text.”
“In my solution to the exam question… is that correct?”
Wide range of academic and practical use cases
Beyond revision and writing support, the prompts cover a broad range of disciplines and use cases, from technical subjects to everyday problem-solving.
Examples include:
“Can you calculate the integral of cos of the third root of x… but only solve the first step of the substitution for me and explain it in an understandable way?”
“Calculate Bayes' rule in this example and explain it in very simple, easy-to-understand terms.”
“Explain the greedy heuristic to me using task 3 as an example.”
Students also use ChatGPT to support applied and professional knowledge:
“Give me a formula in Excel for a complex VLOOKUP or INDEX-MATCH nesting.”
“Explain the basics of a DCF model.”
“Explain the structure of bull and bear spreads.”
Other prompts move into broader academic, legal, and economic reasoning:
“If a representative fraudulently deceives according to § 164, is this considered deception of the contracting party or deception of a third party?
“Under the assumptions of Modigliani and Miller… does this also apply to the return?”
“What is the difference between shareholder and stakeholder?”
The dataset also includes more general and practical requests, such as:
“Which everyday routines can be easily automated and how can I implement this easily?”
“Make me a recipe out of the ingredients I have… I don’t need to use everything.”
Shift in how students use AI
In her post, Bothur argues that the findings challenge assumptions about AI reducing student effort or weakening thinking: “Students are not using AI to avoid thinking. They’re using it to think better!! They are structuring their thoughts, stress-testing arguments, simulating real exam situations & translating complexity into clarity.”
She adds that the intent behind many prompts is to improve work rather than outsource it: “Some students explicitly don’t want the answer rewritten. They ask for feedback, not for AI to do it all FOR THEM. Because they want to be challenged.”
Bothur also frames the examples as evidence of a broader shift already underway: “We keep debating how AI will change education. But we don’t need to speculate anymore. 50 students shared their real OpenAI ChatGPT conversations giving a raw look at how AI is already used for learning. And what they show is not the future of learning, it’s already happening.”
The conversations suggest that students are using AI as part of an iterative study process, particularly for revision, explanation, and structured problem-solving, rather than relying on it to produce final outputs.
