Text this: Evaluating the Performance of AI Large Language Models in Detecting Pediatric Medication Errors Across Languages