Chaotic and unstructured traffic environments, especially in countries like India, pose unique challenges to road safety and autonomy. Dense and diverse traffic, weak infrastructure, lack of delineated lanes, and the presence of pedestrians and animals create conditions that are difficult to manage and yet critical to address. With one of the largest and fastest-growing road networks in the world, improving safety has become more urgent than ever before. Artificial intelligence and machine learning offer new ways to tackle these challenges by studying driver and pedestrian attention, intention, and behavior in complex real-world settings. Recent efforts include the development of datasets and benchmarks that capture unstructured driving scenarios, along with large-scale deployments such as iRASTE, which reduced accidents through the proactive identification of blackspots. Such advances open new possibilities for explainable AI systems that can educate, assess, and assist road users, while addressing the unique safety demands of unstructured driving environments. Two other key aspects of road safety emerge from the attention of the humans (drivers and pedestrians) while driving. In this talk, we will discuss some of the research directions that could help in understanding road safety from a driver's perspective.