Applications like computer vision caused the development of systems, based on methods and algorithms that could recognize something and execute an action based on the result. These new capabilities revolve around real-time situational awareness and automated analysis of very large volumes of sensor data. As a result, technology has moved beyond just proposing task solutions — such as executing a work order or a sales order — to sensing what is happening in the world around it, analyzing that new information for patterns, risks and possibilities, presenting alternatives, and automatically taking actions.
These combinatorial innovations are what we call “complex adaptive systems” where multiple technologies converge and reinforce one another.
For example, the development of autonomous vehicles not only depends on advancements in robotics and artificial intelligence to operate vehicles, but also on the maturation of the Internet of Things so an array of sensors can analyze driving conditions and interact with other cars, as well as improvements in lithium and battery technology for cars to be able to efficiently refuel themselves. The interdependence of these technologies has no doubt contributed to their synchronous advancement. For example, many believed a limiting factor in the emergence of driverless cars was the high cost of batteries required to travel long distances. In response, however, battery producers have dramatically increased production to scale down per unit costs. Over the last ten years, EV battery prices have fallen 90%.
From seemingly disparate innovations powered by self-organizing sensor networks processing voluminous amounts of data, computers are able for the first time to understand and form associations based on statistical methods. Computers can now do what we previously thought only humans could do.
Complex adaptive and autonomous systems have common attributes, including:
- The ability to infer and reason, using substantial amounts of appropriately represented knowledge.
- The ability to learn from their experiences and improve their performance over time.
- The capability of explaining themselves and taking naturally expressed directions.
- The awareness of themselves and ability to reflect on their own behavior and to respond robustly to surprises.
As these systems evolve, we are setting the stage for numerous “invisible” computer-to-computer interactions (a new generation of “machine-to-machine” transactions):
- Autonomous machines making decisions about purchases on our behalf or concluding contracts.
- Multi-agent systems and decentralized autonomous machines enabled to lease themselves out, hire maintenance professionals, and pay for replacement parts.
- Micro-payments between machines, such as a car looking for a specific spare part, or drones and farming systems negotiating directly with each other for services.
Machine data models and analytics allows not only data patterns but a much higher order of intelligence to emerge from large collections of ordinary machine and device data, similar to the neurons of the brain, ants in an anthill, human beings in a society, or information devices connected to each other. The many “nodes” of a network may not be very “smart” in themselves, but if they are networked in a way that allows them to connect effortlessly and interoperate seamlessly, they begin to give rise to complex, system-wide behavior that usually goes by the name “emergence.” That is, an entirely new order of intelligence “emerges” from the system as a whole—an intelligence that could not have been predicted by looking at any of the nodes individually.