Like all technological advances, the most advanced technological tool of our time, deep learning generative artificial intelligence, can ha...
Like all technological advances, the most advanced technological tool of our time, deep learning generative artificial intelligence, can have many beneficial and many dangerous or seemingly dangerous effects on humanity. Often, both the positive and negative effects of a technology can be traced back to the same source in a very recognizable way, and it is really only the human application that shapes the usefulness of how a technology is used.
The development of artificial intelligence is inherently motivated by the need to be a positive facilitator of human activities. The direct social implication of the application of AI is that this development will necessarily make many human activities redundant in the way they are performed by us without AI. Many human activities can already be performed by machines controlled by artificial intelligence, and thus in these areas human activities, and thus even humans themselves, may become redundant. At the same time, however, this development can also be seen as freeing humans from activities that machines can do for them, and typically do better.
Human evolution has followed this development pattern utilizing technology from the beginning of its history. However, the current development differs from the previous ones in that deep learning AI can replace humans in cognitive activities, and this kind of development has never happened before in human history.
The role of government, which coordinates the organization of society, is to manage society properly, and a capable government would also apply advances in AI technology in ways that benefit society. Such a positive social adaptation could be, for example, the provision of a basic income to members of society, while at the same time the government could encourage the creative activities of citizens made possible by the individual and collective well-being created by the application of AI, and support the creation of new things.
Of course, an incompetent government would use the technological advances created by artificial intelligence primarily for the benefit of the ruling elite, both to increase their own wealth and to maintain their rule. In this case, the social degradation of society will obviously follow. This type of AI adaptation is more likely if the current methods of social governance are maintained.
Another concern is that artificial intelligence, advancing through its own evolutionary development, will one day surpass human cognitive abilities in all areas, create artificial general intelligence on its own, and eventually completely replace humans in the field of the evolution of life of increasing complexity, making the human species redundant in the evolutionary development of life and intelligence.
While this is potentially possible, there are several substantive aspects of this development that could have a fundamental impact on such an evolution.
Natural evolutionary development, i.e. development based on diversity and selection for fitness in a given environment, is also potentially possible for the development of AI systems. In special cases, computer systems are already evolving using evolutionary methods.
It is important to note, however, that the evolutionary development of AI typically takes place as AI systems evolve, for example, in competition with each other, while the evolution of capabilities occurs essentially through the organization and use of accumulated data, not by changing the computational code that implements the functions. For example, if two artificial intelligences play chess against each other and soon they could outperform the best human player, the evolutionary process based on learning to play chess does not change the computational code that organizes the data, but it changes the organization of the information collected in the chess games.
Evolutionary mechanisms that increase the efficiency of computational code are also used in computer science, but such advances in artificial intelligence face efficiency limitations because it would require new and new generations of knowledge by the evolved computational code to verify its properly enhanced capabilities.
Moreover, even the most advanced generative deep learning AI currently lacks many of the cognitive functions that are specific to humans, and we do not even know the exact mechanisms by which they are generated in the human brain. Such cognitive functions include the formation of meaning, intention, autonomous volition, and self-consciousness. The realization of these functions is certainly not only the result of the organization of knowledge, of the collected environmental information, but they are also the consequence of the functioning of the controlling code of how to organize the data. In this sense, the natural evolutionary development of artificial intelligence is not an efficient way to develop new capabilities created by computational codes, and human-led evolutionary development of AI is subject to theoretical constraints limited by actual human knowledge.
The most likely scenario for the relationship between human and machine intelligence is a kind of symbiotic coexistence in the future. Humans will be increasingly replaced or, more correctly viewed, supported and complemented by evolving AI, while the consequence of symbiotic coexistence will be an increase in joint potential measured in terms of capabilities and performance. The symbiotic coexistence of humans and AI is the most mutually beneficial and therefore the most likely path for the evolution of intelligent life.
And a less recognized area of impact of this coexistence, which has already begun, is the preservation of human knowledge by artificial intelligence.
In the beginning, the knowledge of mankind could be possessed by a single person. Then, as the accumulated knowledge grew, more and more people were required to possess, use, and transmit all known knowledge, while the personal knowledge possessed by individuals became more specialized. The advent of writing made it possible to preserve knowledge without the need for human existence to carry it, and the formation of libraries made it possible to accumulate knowledge beyond individual human capacity while preserving that knowledge in a passive form. Today, no one can know everything, and more and more specific knowledge and the ways to apply it are held by a few different people.
However, there are fundamental risks to this development. An event of our own making, or an outside influence, could wipe out the existence of people who possess such specific knowledge. As a consequence, actively used knowledge could easily be lost, and all of humanity could easily be thrown back, even stuck for an extended period of time in a significantly earlier stage of its technological and thus social evolution. Humanity is becoming increasingly sensitive to such risks as it evolves.
Generative deep learning artificial intelligence can play an important role in overcoming this kind of risk, because artificial intelligence can not only store and organize the accumulated knowledge, but also reflect the collected knowledge back in a usable form. It can preserve human knowledge in a form that is easily accessible to us. Artificial intelligence makes humanity more sustainable against catastrophic events dangerous to humanity by being able to preserve human knowledge.
Artificial intelligence can be not only a threat or a help to humanity, but also a savior of humanity, for example, by the way of preserving human knowledge in an even more advanced, more active, more usable way than simple writings and libraries of collected knowledge exist. Our coexistence with artificial intelligence can also function as our security backup of our acquired knowledge about the world and about how we can maintain our life in this world.
No comments