Prof. Dr. Murat Sarı was invited as a Plenary Speaker at the "International Conference on Mathematics and Applied Data Science (ICMADS)," which took place at Necmettin Erbakan University in Konya, Turkey, between 29-31 August 2025.

This talk explores how modern artificial intelligence can be used to model real behaviours in complex physical, biological, and socio-economic systems. Moving beyond costly traditional simulations, it presents AI-based surrogates that approximate nonlinear and steep dynamics, drawing on recent case studies and highlighting both the promise and open challenges of data-driven computational science.

This presentation examines how AI-based models can complement and, in some cases, replace traditional numerical approaches for predicting complex physical and biological processes. Starting from the classical notion of modelling as a bridge between theory, experiments and data, it addresses how steep, nonlinear and stochastic behaviours challenge standard partial differential equation solvers. An AI modelling framework is outlined that uses real or simulated data to train neural-network surrogates, validate them against physical models and extend them to more realistic scenarios. Without focusing on a single case, the presentation brings together examples from epidemiology, environmental pollution, agriculture, biomechanics and biomedical applications. Across these domains, AI techniques provide flexible, efficient tools for forecasting and scenario analysis, while also raising important questions about interpretability, data quality and generalization. The conclusion highlights that AI does not remove the need for mathematical insight; rather, it opens new possibilities for hybrid physics–AI approaches and large-scale simulations that would be prohibitively expensive with classical methods alone, underscoring an ongoing paradigm shift in computational science towards data-driven yet physically informed models.