Machines have no dreams
Fancy concerns about artificial intelligence as a threat to human freedom belong to the sort of speculations that various conspiracy theorists appreciate very much. Being a popular topic of numerous discussions on media, AI, nevertheless, remains a blind spot in public awareness. And we are afraid of what we do not understand.
Artificial intelligence is an advanced technology of data processing. This is the best tool for working with ample statistical information. The more data is involved in processing, the better results are achieved finally. The Big Data term belongs to those petabytes of raw information that programmers upload to their AI algorithms. Without Big Data any AI solution is useless. What frightens us so in it then?
Perhaps, we find some practical use cases of AI scary. Sometimes, manifestations and effects of AI look so humanly. Face recognition in our smartphones is quite intimidating with regard to its evolving accurateness. Voice assistants start talking to us so naturally. Chatbots on social media behave oftentimes so rationally that it is almost impossible to distinguish them from human users. Namely human-like features of AI make many of those who know nothing about the technology rattled.
You may argue that there are some true professionals among the AI fearmongers, the ones who cannot be suspected in ignorance. They know what they are talking about. Elon Musk, for example, can barely rely on his warnings on just amateurish superstition. His autonomous Tesla trucks are full of AI algorithms. But it’s hard to say he is lying as well.
Undoubtedly, Elon tells the truth. But the thing is that the majority who listen to him are unable to distinguish an actual danger of AI from popular flimsy AI bugaboos. He warns about AI in terms of its unprecedented efficiency when AI as a super effective instrument that might appear in the service of some evil forces. This is not about machines’ cruel subjugation of humanity. AI is neither good nor evil in itself since even the most profound AI system cannot have any independent will.
Any volition is impossible without a desire while any desire is based on emotions at its core. What is an emotion? Nothing but a mistake, a software glitch, a bug, a contradiction between what takes place and what is to be. Anger, sorrow, jealousy, envy, as well as all the other human emotions, cannot be pre-programmed within any AI. Know why? Because the only objective of AI is to minimize mistakes.
Avoiding mistakes through strictly following the rules is just what we call intelligence, right? This is why playing chess (or Go) and driving cars fit the capabilities of the advanced AI systems very well. AI is self-educating because it follows the pre-programmed algorithms when processing Big Data. The very effectiveness of the AI-based solutions is grounded on a faultless modus operandi inherent in systems running on algorithms. The famous “If This Then That” programming principle does not allow any equivocal interpretations. If you voluntary change even one rule in chess, the most powerful chess-playing AI such as Deep Blue will freeze.
Moreover, any AI cannot break the rules in the algorithms it runs on — this is simply impossible in technical terms. This is what will keep machines always behind humans. Even the most intelligent computers can never create anything utterly new, anything going beyond their inbuilt algorithms.
Some may argue that even available AIs can create paintings competing with the art of some French impressionists. It just seems so. In fact, what we see as new paintings created by AI are more or less successful compilations and variations on the theme. The same relates to the “new” music and literature created by AI. It is us humans who can recognize and evaluate the “art” of AI with our human imagination. No AI is capable of criticizing the piecework of another AI — they both lack consciousnesses to do that. Imagination appears only if a creature possesses a self-awareness. And the first mental category distinguishing humans from both the other living beings and intelligent machines is time.
Neither animals nor machines are able to imagine an event happening not at the current moment. Time is so hard to comprehend that we humans are still confused with a final definition of time. Time is unknowable without a critical capability of distinguishing “me” from “all the others”. The past and the future are extremely difficult complex concepts to be explained.
Our mind operates with those complex concepts while AI operates with primitive inputs even though the inputs are collected in complex patterns. No creativity will appear even if you combine multiple complex patterns built on simple commands. No matter what computing power is provided for one or another AI — a predefined logic of programming languages will always determine horizons of the possible for machines.
If we look at the history of humankind from a machine’s perspective, very few logic can be found in the events that changed the course of development of our civilization. One of the most prominent examples of them is the emergence of Christianity. “Credo quia absurdum” — I believe because it’s absurd, as Thomas Aquinas said.
Hundreds of years Christians were staying against the very paradigm of the pre-Christian world once their new belief was nothing but a pure paradox for the dominating religions those days. Namely, the paradoxical ideas of Jesus Christ constituted that historic turning point which changed the entire civilizational paradigm of the whole humanity for millennia. Nothing but a critical bug Christianity could be for AI since no rational explanations were inherent in how first Christians lived and died.
Machines cannot believe by definition. They also cannot suspect, doubt, hope, wish and, therefore, no epiphany is possible for any AI. No matter how powerful and intelligent a machine is, it will always remain an instrument through which some human will can be provided. That was just what Elon Musk warned against with regard to a danger that could be caused by a super AI to humanity. Not advanced algorithms but evil intentions are dangerous. And intentions, as we see, are inherent in humans, not in machines.
Thus, we should live AI algorithms alone as well as their programmers — they simply do what they are paid for. It is better to keep a close eye on who hires the AI programmers and for what. But here we won’t do without conspiracy.