Some serious issue with Backpropagation and Deep Learning in the Neural Network Analysis video

I'm no expert in ANNs, but I have worked with them for over 10 years + I'm involved in the development of a course on AI. To my understanding, backpropagation is a training algorithm for all kinds of ANN, making use of 1. errors in the predictions of the ANN and 2. the derivatives of the error function. The aim is to provide appropriate corrections to each one of the neurons involved in the ANN.  On the other hand, Deep Learning (DL) is the methodology involving large-scale ANN (usually consisting of many layers of neurons), sometimes highly specialized for specific tasks (e.g. voice recognition, image processing, and video analysis). From this, we can conclude that the statements made in the last part of this video are utterly inaccurate. Also, it's unclear from this video what the business value of these tools is (although I have no doubt that they are very valuable, especially in areas like Healthcare, Telecoms, and Manufacturing). Cheers