32 Case Study: Comparing uses
The scenario
Yasmine is working on a statistics project for her sociology class, investigating if there’s a connection between students’ participation in school clubs and their reported stress levels. She has collected survey data from 200 classmates. Yasmine first uses statistical software to organize her data. She then opens a GenAI tool and asks, “What are common statistical tests to compare two groups of data, and what are the main assumptions for each?” The GenAI gives explanations for tests like the t-test and ANOVA. Yasmine reviews these, picks the most suitable test based on her data type, runs the analysis herself in the statistical software, and then interprets the actual results from her software output. She might use the GenAI again to clarify a specific term in the output, like “p-value,” ensuring she accurately understands her own findings.
David is also working on a similar statistics project but feels overwhelmed by the data analysis. Instead of trying to understand the process, they upload their raw survey data (containing sensitive student responses) directly to a GenAI tool. They prompt it: “Analyze this data about student club participation and stress levels, and write a full report, including all relevant graphs and interpretations, for my statistics project.” The GenAI quickly generates a report with charts, statistical terms, and conclusions. David, without checking the accuracy of the graphs or truly understanding the statistical methods used, copies and pastes the AI-generated report directly into their project, presenting it as their own analysis and interpretation. They don’t take the time to verify if the GenAI made up any “insights” or if its conclusions even logically follow the actual data.
How do these situations compare?
- Yasmine uses GenAI as a tool for information and understanding, not as a substitute for her own work including
- Seeking guidance: Yasmine asks the GenAI what statistical tests are common for her type of data and their assumptions. This is like asking a subject matter expert for a list of potential methods.
- Clarifying concepts: Yasmine uses the GenAI to understand a specific term, like “p-value.” This helps her better interpret the results she generated herself, promoting learning.
- Aiding decision-making: The GenAI’s output helps her decide which test is most suitable, but she performs the actual analysis herself in the statistical software.
- David’s approach to using GenAI is academically inappropriate, unethical, and demonstrates a failure to engage with the assignment including
- Outsourcing the entire task: David asks the GenAI to perform the entire analysis, including creating graphs, interpreting results, and writing the final report. This is outsourcing the core learning objectives of the project, which are to analyze data and interpret findings.
- Compromising data privacy: David uploads sensitive student data directly to a GenAI tool. This is a significant breach of privacy and a major ethical concern.
- Ignoring verification: David does not check the accuracy of the AI-generated report. This demonstrates a lack of critical thinking and a willingness to submit work without ensuring it’s correct or based on reality. The AI could have hallucinated data or generated flawed conclusions.
