New Perspectives in Science Education

The rapid evolution of generative artificial intelligence (GenAI) has challenged traditional inquiry-based models, such as the original WebQuest launched in 1995. While the initial framework aimed to scaffold cognition by pre-selecting web resources, modern GenAI tools often bypass the reflective process, leading to a "resolution gap" where students accept automated answers without critical transformation. This paper proposes "WebQuest 3.0," an updated methodological framework designed for the 2026 educational horizon. WebQuest 3.0 redefines the AI’s role from a simple information provider to a "Socratic Partner." Within this framework, scientific tasks are "AI-proofed" by requiring students to audit, compare, and verify AI-generated hypotheses against Open Educational Resources (RED). The study aligns this approach with the TPCK framework and the European Digital Competence Frameworks (DigComp 2.2 and DigCompEdu), addressing specific implementation challenges within the Romanian educational context. Preliminary findings suggest that when AI is integrated through structured "interaction circuits" and socratic prompting, students demonstrate higher levels of metacognitive engagement and a more robust understanding of the scientific method. The paper provides practical prototypes for science educators, offering a rubric to evaluate the quality of human-AI dialogue, thus ensuring that technology serves as an e-scaffold for independent critical thinking rather than a substitute for intellectual effort.