Development of Mobile Applications Based on Artificial Intelligence: Current Experience and Prospects

This article explores contemporary approaches to the development of mobile applications using artificial intelligence (AI) technologies. It presents a comparative analysis of available tools and frameworks and outlines key directions for their future advancement. The relevance of this study is driven by the rapid growth in demand for intelligent mobile services and the need to systematize practices that combine automation, personalization, and autonomous data processing. The aim of the research is to summarize current practices and identify prospects for the development of AI-based mobile applications. The study analyzes the capabilities and limitations of popular platforms (Softr, FlutterFlow, Thunkable, Microsoft Power Apps, GitHub Copilot Mobile, AppMySite) based on criteria such as intended use, technical complexity, functionality, scalability, and cost. A structured expert-based evaluation enabled the classification of tools according to the needs of startups, the corporate sector, beginners, and professional developers. The article emphasizes the role of cloud platforms (Google Cloud AI, AWS AI Services, Microsoft Azure AI) and mobile SDKs (TensorFlow Lite, Core ML, ML Kit) that provide the technical foundation for implementing AI models in mobile applications. Key challenges in the current development stage are identified, including the need for high-quality data, limited computational resources, model optimization for mobile environments, privacy risks, and ethical dilemmas, particularly algorithmic bias. Promising areas for further development are outlined, such as the growing importance of Edge AI, the implementation of explainable AI (XAI), hyper-personalization of user experience, enhanced cybersecurity through intelligent systems, the emergence of new app categories based on generative AI, and the democratization of development through low-code/no-code environments. General conclusions are offered regarding the effectiveness of different tools depending on project conditions and user expertise. The findings can be applied in the improvement of software engineering curricula, digital service development in business and public administration, and in selecting strategies for integrating AI into mobile products.