Trusting Minds vs. Machines

Overview

Objective: This project investigated whether humans find answers to science questions more trustworthy when provided by an artificial intelligence system or a human. As AI becomes a tool for schoolwork and everyday decisions, it is important to understand how it affects our thinking. My goal was to determine if AI can successfully mimic human written expertise in a way that gains equal or greater public trust. Methods: To assess this, I first considered research from the MIT Media Lab where EEG brain scans showed that relying too much on AI can lower brain activity by up to 55 percent. This creates a state called cognitive debt where our brains might not work as hard. To test if this affects trust, I selected ten diverse science topics and generated two matched explanations for each using ChatGPT and one written by a human. I ensured both responses were similar in length and clarity to maintain fairness. I then distributed a blind study using a Google Form to 213 participants, asking them to identify which answer they trusted most and to guess the author of each response. Results: The findings reveal a significant illusion of human preference. Participants frequently selected the AI generated answer as more trustworthy, even when they firmly believed they were choosing a human author. Data indicates that when participants attempted to identify human writing, they actually chose the AI generated response 59 percent of the time. Conclusion: My hypothesis was supported because the data demonstrates that modern AI content is often indistinguishable from human writing. While participants claimed a preference for human expertise, they were unable to accurately identify the source. These findings suggest that as AI becomes more sophisticated, we must remain vigilant in evaluating our information sources to maintain strong critical thinking skills.

Competition history

• CSEF 2026 Cognitive Science (Junior Division) · Entry J-06-12

Related projects

Closest projects by meaning, across every fair and year in the corpus.

Browse more like this

• All Behavioural and Social Sciences projects
• CSEF Behavioural and Social Sciences projects
• Every CSEF 2026 project