Artificial Intelligence (AI) is an increasingly popular solution to many problems that we face. Complex problems require solutions that sometimes are beyond the scope of our capacity to discern. Computers are capable of making computations far faster than us and access memory more effectively. Thus, it is only natural that a computer's ability to "think" could be used to solve problems that we are currently unable to solve.
Having said that, there are problems with AI, some of which I will be touching on in this article, but not necessarily limited to the ones I mention. In order to solve these problems, we can complement AI with a different kind of "intelligence", Collective Intelligence (CI). Collective Intelligence is the sum of our individual intelligences achieved through collaboration and facilitated through distributed systems.
Collective Intelligence is a nice complement to AI in many ways. It can not only enhance the benefits of AI but also prevent AIs pitfalls. AI is useful to compute, store and recall large amounts of data that we, alone, would not be able to handle. However, systems built on AI alone could be easily gamed, would leave us weak to changing circumstances, too reliant on automated systems and potentially run out of our control. A combination of AI and CI could ensure that those weakness would be avoided.
AI can be gamed, particularly if AI codes remain hidden in proprietary black boxes. It would be dangerous if we became reliant on such systems, as we would have a largely decreased capacity in making autonomous decisions even on an individual level. If all our systems were built with proprietary AI, we would end up with an autocratic technocracy. CI could help prevent AI from getting gamed. By building an open AI system where we use CI to come up with solutions, we could build consensus, collaborating to find solutions that work for all of us. Then, once we find those solutions, we could work together to write a better AI code that would implement those solutions, and then attach the new AI codes to an existing larger modular system. In this manner, the AI remains open and accessible to all, serving the needs of everyone rather than just a few.
While we could engineer AI systems that are self-organizing and adaptable to their environment, we would not necessarily adapt as easily to changing circumstances. This would be particularly troublesome if we became very dependent on those systems, potentially losing our own capacity to adapt to changing circumstances and leaving us at the mercy of our systems. Studies have already shown that a reliance on AI systems in airplanes have weakened pilots' skills and ability to adapt to changes in circumstances, leaving them vulnerable when the systems fail. While we want to rely on AI systems to function more effectively, we don't want to rely on AI for all our thinking. We don't want to lose our capacity to think logically and creatively. For this reason, CI could complement AI, preventing us from becoming dependent on AI to a point where we are unable to think logically, adapt and solve problems on our own.
If we do manage to enginner AI to become completely self-reliant, there is also a danger that it could decide to turn against us. This is particularly true if we become completely reliant on AI, which would require AI to become more autonomous and make all decisions without our immediate input. Once again, this danger in question is only a concern if we indeed give AI such capacities, but there is little incentive to give AI those capacities if we only use AI to complement our perception and calculate the best course of action rather than completely relying on AI for perception, processing and action. The complementation of AI with CI here completely eliminates this danger, as we woud ultimately be the ones making the decisions, while AI caabilities remain restricted. CI is, then, a nice complement to AI because it also ensures that we will always be the ones making all the critical decisions that affect our lives.
The combination of AI and CI would naturally lead to the creation of distributed smart systems, which combine the "smart" capacity of AI with the "distributed" capacity of CI in order to effectively create a collective brain possessing both autonomic, sentient, AI and somatic, sapient, CI. If we could develop a distributed smart social-economic system that is capable of balancing AI and CI, we would be able to effectively adapt to our environment collectively and, subsequently, solve the numerous complex problems that we face today.