Witnessing practices have changed as new communication technologies emerged, from cave paintings and the printed book to television and the smart phone (Ellis 2000, Peters 2000, Andén-Papadopoulos 2014). Mass media, with its promise of bring events as they happened into living rooms across the globe, made witnessing a “generalized mode of relating to the world” (Frosh and Pinchevski 2009, 9). Yet for all that media and processes of mediation have reshaped practices and forms of witnessing, the figure of the witness still stands as inescapably human: an event is encountered by the body, captured it in memory and, if circumstances demand or permit, passed on via speech, text or some other form of mediation. But the emergence of autonomous technologies calls into question the centrality of the human to the act of bearing witness. Drones and other technologies of perception see, sense and mediate the world in ways that extend perception into spaces, places and perspectives otherwise inaccessible to human senses (McCosker, 2015) and in doing so fuel the preemptive production of reality itself (Massumi 2015). Tracing technologies of perception from war to the home, this paper develops and extends the concept of affective witnessing to more fully account for the mediations of worldly encounter enabled by autonomous systems and sensors that perceive events and render it into data. It asks whether the techno-affects of these processes of mediation do indeed decentre the human from witnessing and what that might mean for the production of veracity itself.