Abstract
Background/Objectives: Periodontitis is the sixth most prevalent disease worldwide, affecting over 740 million people and representing a growing public health burden, particularly in transitional economies. While artificial intelligence (AI) technologies -including deep learning (DL) and machine learning (ML) -have demonstrated high diagnostic accuracy in controlled research settings, a critical implementation gap persists between proof-of-concept studies and real-world clinical deployment, especially in countries with nascent regulatory frameworks such as Kazakhstan. This systematic scoping review synthesizes global evidence on AI methodologies applied to periodontal diagnosis, risk prediction, and patient monitoring; evaluates regulatory and governance frameworks; and proposes an evidence-based five-stage implementation model contextualized for emerging health systems. Methods: Following PRISMA 2020 guidelines, a systematic literature search was conducted in PubMed/MEDLINE, Scopus, Web of Science, Embase, and Cochrane Library, supplemented by regulatory and grey literature. Search terms combined (‘artificial intelligence’ OR ‘machine learning’ OR ‘deep learning’) AND (‘periodontitis’ OR ‘periodontal disease’ OR ‘gingivitis’). Studies published from January 2015 to April 2025 were considered. Forty-five sources (36 peer-reviewed empirical and review studies plus 9 regulatory and grey-literature documents) meeting predefined inclusion criteria were included in qualitative synthesis. Results: Two dominant AI paradigms were identified: (1) image-based deep learning -particularly convolutional neural networks (CNNs) and Vision Transformers -achieving diagnostic accuracies of 73–98.6% for periodontal bone loss detection on panoramic and periapical radiographs; and (2) ML-based non-clinical predictive screening models using patient-reported data and salivary biomarkers. Digital patient management tools (smart toothbrushes, chatbots, IoT platforms) represent a third emerging application domain. Significant external validation performance drops were consistently observed across models, underscoring data quality, dataset heterogeneity, and sample-size constraints as primary barriers. Regulatory analysis revealed convergence of the EU AI Act, U.S. FDA framework, WHO guidance, and the emerging Kazakhstan AI Development Concept (2024–2029) around risk-based classification, algorithmic transparency, and post-market surveillance mandates. Conclusions: Safe, ethical, and clinically effective integration of AI in periodontology requires a phased implementation approach: establishing national multimodal databases, performing locally validated clinical trials, integrating certified tools into digital workflows, deploying continuous monitoring systems, and developing robust legal governance including liability and insurance mechanisms. Kazakhstan’s rapidly evolving regulatory infrastructure and digitalization strategy create a concrete opportunity to establish a replicable model for AI adoption in periodontal care across Central Asian and transitional economies.