Third-party cyber evaluations involving OpenAI models
openai.com·14h ago
AI Research
arXiv CS.AI·Jul 2·5 min read
Constructing Epistemic AI Literacy: Detecting Epistemic Aims and Processes in Student-AI Co-Programming
TL;DR
Students often struggle to effectively collaborate with AI in programming tasks due to a lack of understanding of epistemic aims, which are the goals related to knowledge acquisition. A framework was developed to detect and analyze these epistemic aims and processes during student-AI co-programming sessions.
✦ Why It Matters
Engineers and researchers can improve AI collaboration by integrating epistemic literacy training into programming education.
Key Takeaways
- Incorporate EAIL framework in AI education.
- Focus on mastery-oriented aims in student interactions.
- Encourage advanced epistemic strategies like justification.
- Evaluate and adapt teaching methods based on EAIL findings.
Full Summary
Epistemic thinking is crucial for students learning to use generative artificial intelligence (GenAI) in programming, as it involves constructing queries and evaluating AI outputs. The study introduces the Epistemic AI Literacy (EAIL) framework, which views AI literacy as a dynamic process shaped by human-AI interactions.
Utilizing the AIR framework, the research analyzes a large dataset of dialogues from student-GenAI co-programming activities to identify observable dimensions of epistemic aims and processes. Results show that 78.8% of interactions were driven by non-mastery-oriented aims, employing less reliable strategies like outsourcing and verification-seeking.
Only 11.1% of interactions demonstrated high epistemic engagement, characterized by mastery-oriented aims and advanced strategies such as epistemic justification. These findings highlight a significant gap in effective AI literacy among students, suggesting a need for improved educational approaches.