AI-Driven APIs: Enabling Adaptive Interoperability in Complex Systems

Authors: Sergio Duarte, Tarik Eltaeib

SUNY Campus: Farmingdale State College

Presentation Type: Poster

Location: Old Union Hall

Presentation #: 13

Timeslot: Session C 1:45-2:45 PM

Abstract: Modern technology ecosystems struggle with interoperability due to the variety of devices, protocols, and data formats. This project investigates how artificial intelligence (AI) enhances application programming interfaces (APIs) to automate communication between systems. Traditional APIs allow systems to share data but often fail to adapt to evolving standards or semantic differences. By integrating AI such as machine learning (ML) and natural language processing (NLP) APIs can automatically adapt to different protocols, resolve data mismatches, and predict integration needs. For example, AI models trained on API usage patterns dynamically translate formats such as JSON to XML and detect security risks. Case studies in IoT and smart manufacturing demonstrate AI-augmented APIs cut manual setup by 40–60% and enhance real-time decisions. Methods include using neural networks to translate data formats and reinforcement learning to manage data flow efficiently. Results show AI enables APIs to evolve with system updates, reducing reliance on static standards. Challenges like limited data and high computing costs are addressed through federated learning to protect privacy. This work highlights AI’s role in building scalable, self-adapting systems, with applications in healthcare, smart cities, and Industry 4.0. Future work focuses on ethical guidelines for autonomous APIs and energy-efficient AI models for edge devices. By combining AI’s flexibility with APIs’ structure, this approach moves beyond rigid standards, ensuring resilient, future-ready interoperability.