Learners’ ability to detect errors in AI-generated content and its impact on critical thinking

Ascertain the extent to which learners can detect inaccuracies in AI-generated content and determine whether learning with ChatGPT fosters critical thinking or reinforces misconceptions in physics education contexts.

Background

The authors discuss that LLMs often present information persuasively, which can mask inaccuracies and may lead students to uncritically accept responses. Prior studies show substantial trust in ChatGPT among physics students, raising concerns about error detection and conceptual understanding.

Clarifying whether learning with ChatGPT enhances critical thinking or instead entrenches misconceptions is essential for safe and effective integration in classrooms.

References

At this point, it is unclear to what extent learners are able to identify errors in AI-generated content and whether learning with ChatGPT fosters critical thinking or, conversely, reinforces misconceptions [5].

— Building Bridges: AI Custom Chatbots as Mediators between Mathematics and Physics (2412.15747 - Lademann et al., 20 Dec 2024) in Section II.B.1 (AI Chatbots in Education: Learning Experience and Learning Performance)