<![CDATA[ABSTRACT This study aims to examine efforts to optimize MIG welding technique learning through the use of artificial intelligence (AI). This study compares three platforms: Monica, Claude, and Perplexity. The results were then evaluated using expert judgment. The background to this research stems from the difficulty students face in recognizing various types of defects in welding results, which results in decreased joint quality and a poor understanding of the welding process as a whole. This study employed a comparative experimental method involving expert validators to validate the analysis results. Three images of MIG welding defects were analyzed separately by the three AI platforms, and the identification results were validated by welding experts. Based on the findings, Perplexity proved to have the highest accuracy in identifying defects, followed by Monica, while Claude performed the lowest. These findings indicate that AI integration can accelerate the identification process, simplify analysis, and make a positive contribution to welding technique learning. However, the study also revealed limitations, particularly related to the use of complex technical language and local contexts that AI has not yet fully accommodated. Therefore, it can be concluded that the application of AI, specifically Perplexity, has the potential to be an effective innovation to support the welding learning process in educational settings.]]>
Copyrights © 2025
