Children's Overtrust and Shifting Perspectives of Generative AI
Project Overview
The document examines the role of generative AI (genAI) in education, focusing on middle school girls' perceptions and experiences with technologies like ChatGPT. Initially, these students exhibit a strong trust in genAI, which transitions to skepticism as they encounter its limitations and the potential for misinformation and bias. This shift in perception emphasizes the critical need for AI literacy within educational settings, enabling students to better navigate the complexities and challenges associated with generative AI. The findings advocate for a deeper understanding of children's perspectives to shape effective educational policies and practices surrounding the integration of genAI in schools. Overall, the research highlights both the potential benefits and the pitfalls of using generative AI in education, stressing the importance of preparing students to engage with these technologies responsibly and critically.
Key Applications
ChatGPT and generative AI tools
Context: Middle school girls participating in STEM outreach programs
Implementation: Workshops integrating discussions and activities focused on genAI limitations, capabilities, and bias
Outcomes: Students shifted from overtrusting genAI to a more critical understanding of its limitations and potential biases; expressed nuanced views on school policies regarding genAI usage
Challenges: Initial overtrust in genAI outputs; difficulty in distinguishing between AI and conventional computational tools; concerns about misinformation and academic integrity
Implementation Barriers
Perceptual Barrier
Students initially overestimate the capabilities of genAI, leading to overtrust in its outputs.
Proposed Solutions: Exposure to incorrect outputs and discussions about limitations helped shift perceptions and fostered critical thinking.
Educational Barrier
Existing media literacy frameworks do not adequately address the unique challenges posed by genAI outputs.
Proposed Solutions: Updating curricula to include specific lessons on evaluating genAI content and its implications.
Project Team
Jaemarie Solyst
Researcher
Ellia Yang
Researcher
Shixian Xie
Researcher
Jessica Hammer
Researcher
Amy Ogan
Researcher
Motahhare Eslami
Researcher
Contact Information
For information about the paper, please contact the authors.
Authors: Jaemarie Solyst, Ellia Yang, Shixian Xie, Jessica Hammer, Amy Ogan, Motahhare Eslami
Source Publication: View Original PaperLink opens in a new window
Project Contact: Dr. Jianhua Yang
LLM Model Version: gpt-4o-mini-2024-07-18
Analysis Provider: Openai