Developing an Interactive Plastic Surgery Education Website Using Manus AI

Abstract

PURPOSE: Agentic AI systems significantly evolve from traditional task-specific AI, offering autonomous reasoning, adaptive learning, and the capacity to orchestrate complex, multi-step tasks. Manus AI, developed in 2025, is among the first of these systems, capable of independently generating, designing, and deploying functional websites. While such technologies are increasingly applied in commercial and educational contexts, their application in healthcare needs further exploration. This study aims to evaluate Manus AIs ability to autonomously develop a comprehensive educational website covering ten plastic surgery procedures. METHODS: A single zero-shot prompt instructed Manus AI to build and deploy an interactive website with structured sections for each procedure, including introduction, indications, risks, recovery, and FAQs. Content quality (accuracy, comprehensiveness, structure, navigation, tone/style/language) was evaluated by six reviewers using a 5-point Likert scale. Usability was assessed via the System Usability Scale (SUS), and readability was measured using the Flesch Reading Ease (FRE), Flesch-Kincaid Grade Level (FKGL) scores, and word count for each procedure page. RESULTS: The AI successfully produced a functional public website with intuitive navigation, a consistent header, and patient-centered design. Content structure, navigation, and tone/style/language achieved the highest mean scores (3.8, 3.8, and 4.1, respectively), while accuracy (3.6) and comprehensiveness (3.5) were lowest, reflecting incomplete detail and occasional informational gaps. Usability scored a mean of 81.25 (SD 7.37), exceeding the 68-point benchmark for above-average performance and ranking in the top 10% of SUS scores, suggesting strong user acceptance. Readability performance was poor (mean FRE 23.76; FKGL 12.7), indicating a college-level reading requirement. Word count inconsistency was also marked, with some pages exceeding 1,000 words and others under 100. Additionally, limitations in figure selection and medical illustration accuracy were identified, indicating possible limitations of current agentic systems in handling complex tasks. CONCLUSION: Manus AI demonstrated the capacity to autonomously generate and deploy aplastic surgery education website, with strong usability and design strengths. However, limitations in accuracy, comprehensiveness, readability, and visuals highlight the need for expert oversight and specialized resources to ensure safe, reliable deployment in healthcare education.© 2026. Plastic Surgery Research Council | All rights reserved |*Source: https://ps-rc.org/meeting/Program/2026/AS51.cgi*

Ähnliche Arbeiten