Systems and methods for creating video watch parties are shown and described. Video and audio data indicative of the content being displayed on a plurality of smart TVs is compared to stored data in a content recognition database to determine what content is being viewed by each of a plurality of viewers at a given time. Viewers watching the same content are invited to become members of a watch party, and the televisions of those who accept the invitation are operatively connected to one another in a peer-to-peer network, enabling them to view the watch party content on a watch party screen that displays the watch party content and images or graphical icon representations of each other member of the watch party who is watching the content. Watch party participants can communicate with one another on the watch party display orally, using text messages in a text message area of the display, or by projecting emojis in the content region of other participants' displays.