Security Vulnerabilities in AI pipeline

The intersection of artificial intelligence and healthcare has introduced promising applications in the early diagnosis of neurodevelopmental disorders such as Autism Spectrum Disorder (ASD). This book presents a web-based AI model for detecting autism using facial images and symptom descriptions, with a strong emphasis on implementing security measures. While the Vision Transformer (ViT) forms the core of the classification engine, the uniqueness of the project lies in safeguarding sensitive health data through advanced encryption, access control, and secure deployment practices. This report explores the methodology, literature background, research objectives, and results with a special focus on data privacy and application.