Oral potentially malignant disorders (OPMDs) are conditions that carry an increased risk of progressing into malignant lesions, including oral cancer. These disorders can manifest in various forms, such as white, red, or mixed red-white lesions, appearing as verrucous, papillary, or ulcerated. Many oral cancers, particularly those linked to habits, often develop from pre-existing OPMDs. Studies indicate that between 0.1% and 40% of leukoplakia cases may progress to oral cancer. Diagnosing OPMDs can be challenging as conditions like frictional keratosis, leukoedema, and candidiasis share similar clinical features. Early identification is crucial to preventing malignant transformation and improving treatment outcomes. Oral cancers share common risk factors and genetic changes with OPMDs, but while the oral cavity is accessible for visual inspection, diagnosis remains subjective and often requires biopsy for confirmation. Technological advances like chemifluorescence, liquid biopsy, and Velscope have improved cancer detection but remain costly and not widely available. Artificial intelligence (AI) has the potential to assist in identifying these lesions through low-cost, non-invasive methods. Studies have applied AI to detect various oral diseases, with promising results, yet practical applications for detecting oral cancer remain scarce. This study aims to develop AI-assisted application to detect OPMDs and oral cancer using photographs. The AI models, including convolutional neural networks (CNNs), aim to identify suspicious lesions. By using AI, the study aims to improve early detection, making screening more efficient, especially for high-risk individuals in areas with limited access to specialized care.