So here's what appears to have been done. The editor took a 16 x 9 aspect ratio video. The man talking. He placed that inside another video with an aspect ratio of 1 x 1. Then he scaled the 16 x 9 video so that it fit from side to side almost exactly. He actually scaled it down a bit smaller than he needed to and added gradient borders to the sides of the video. For esthetics I suspect but it could have been scaled to fit side to side just as easily.
Bottom line, you would need to create a square video to reproduce what your image depicts.