There is a pronounced tendency on the part of most observers of technology-related change today to look for the “next big thing” or the single “disruptive innovation” that will almost single-handedly redefine some economic space or some key aspect of life. The profound and long-term impacts that new technologies have on societies, however, have often had long lag times and required the evolution of some degree of new “ecosystem niches,” either in terms of new skills and models or ancillary supporting technologies (or both). This type of framing of technological change (and the all-important impacts on society and daily life) is probably more helpful when thinking about the futures of machine autonomy than the “next big thing” approach to anticipating change.
I was reminded of this by a Quartz article this morning about the Amazon Echo, a rather unobtrusive device that, as the author of the piece called it, seems to be a sleeper hit. The Echo is voice-interactive and always on – apparently no more frustrating efforts to first open the program (like Now or Cortana on your desktop); the device is always listening and responds to a slowly increasing range of inquiries and requests, from converting measurements to now also being able to call up an Uber ride.
Does Echo signal that Amazon is poised to take over the AI assistant space or that Echo itself will redefine the future of AI and machine autonomy? No. But the growing success of Echo, which slips into the background of your life and becomes a second-nature action is an indication of some elements of an ecosystem of assistant machine autonomy. Echo is a reminder that in order for some of our more ambitious images of a machine future to emerge and become embedded in daily life, a wide range of niches in the man-machine world have to be created and then maintained.
Taking this cue, more useful thinking about the futures of machine autonomy in our human world would think about how a variety of different niches might emerge and then connect to one another. It’s a question of how a machine autonomy ecosystem might emerge and evolve. It’s a context in which there will be a lot of competition for dominating new niches as well as co-evolution between multiple solutions that begin to rely on each other and impact each other’s evolution.In terms of forecasting, there’s no one right answer to the question of “what will this all look like in 20 years.” In today’s world, there are so many people trying to create and fill niches, in so many different aspects of daily life, that the resulting turbulence and uncertainty defies most traditional approaches to forecasting change. But such high levels of turbulence also imply that there is greater opportunity for individuals, organizations, and communities to become active and to shape change, to influence the structures and assumptions about what roles machines play in daily life.