The Standard Model (SM) of particle physics offers a foundation for understanding fundamental particles and their interactions, yet mysteries persist, including dark matter and dark energy. This colloquium talk explores innovative methods to address these gaps, focusing on the ATLAS experiment at the Large Hadron Collider (LHC). I will discuss model-dependent searches aimed at discovering new particles acting as mediators between the dark sector and the SM. Additionally, I will highlight the growing importance of model-independent anomaly detection techniques in our quest for new physics. Integration of innovative data analysis techniques, including Artificial Intelligence (AI) and Machine Learning (ML) methods becoming crucial to address several challenges in searches for physics beyond the SM. Specifically, I will present a novel ML-based approach designed to enhance model-independent anomaly detection capabilities. Finally, I will discuss one of our major experimental challenges coming from the large data volume. It will be even more challenging in the future High-Luminosity LHC era. I will highlight our current data filtering strategy through an experimental “trigger” system to reduce the data volume and how real-time AI algorithms on trigger hardware like Field-programmable Gate Arrays (FPGA) and application-specific integrated circuits (ASIC) could mitigate some of these problems and accelerate discoveries.