Texting while driving poses serious safety risks, as it is 6 times more likely to cause an accident than drunk driving and accounts for over 3,000 deaths per year. Researchers are exploring using computer vision techniques like bag of visual words models and classifiers like SVM, random forest, and XGBoost to detect texting drivers from images. When trained on a dataset of over 10,000 images, these models achieved overall accuracies between 72% to 97% at distinguishing between images of drivers who were texting versus driving normally.