Conscious Computers Not Possible... But AI Is Still Dangerous
- R. Totten (c) 2016, 2018
The notion of actually "conscious" machines (computers) in the realm of artificial intelligence (AI) does not at all promise to be a realistic possibility. This is because actual "intelligence" is necessarily connected to a real consciousness of a sentient living thing --such as a person or an animal. True consciousness is a property of a spirit inside a sentient being.
Some people suppose that computers already "think" and are "intelligent," and to these people this seems to be so, because computers are able to accomplish amazing calculations and logic-processes at tremendous speeds which make virtually all people appear "stupid" in comparison. Many elderly people feel that computers are already conscious and able to reason and think.
--However, such machine "intelligence" only appears to be the case; an illusion. Even AI computers with "super" intelligence would simply be a "super" illusion of highly effective human ingenuity at mimicking intelligence. --This illusion is even now assumed and accepted by observers only because the people making this assumption have an inadequate understanding of what computer information is, and that such "information" or "intelligence" has nothing to do with actual awareness or consciousness on the part of the computer. Uninformed people don't realize that electronic computer information (and "memory") is actually just tiny electronic switches (in the form of transistors or electric charges) that are either turned "on" or turned "off." In the end, these switches are no more profound (with regard to the computer being sentient or conscious) than the much larger switches which turn off and on lights in the rooms of a house. There is no more consciousness involved in such electronic "memory," than there is in having information written on a piece of paper. It is an illusion of "memory," but not actual memory connected to a conscious, reasoning mind.
In computer information and memory, a switch that is "on" is called a "one," but a switch that is "off" is called a "zero." Any piece of data contained in the switches of computer "memory" is actually a lengthy series of these ones and zeros which stands for that string of information. Larger amounts of such "digital information" are stored in longer series of ones and zeros in the transistor-switches of computer memory. Even sound or pictures are just a bunch of ones and zeros. In addition, operations of computer "logic" are also lengthy series of ones and zeros (expressing logarithms) which cause various transistor-switches and "logic-circuits" of the computer's processors and electronic "memory" to respond and behave in particular ways. The computer itself has no actual awareness of any logical necessity in any discussion or argument, but it is the human programmers whose logic is reflected. No computer is actually thinking or aware.
Again, since the switches and electric charges of computer memory are much smaller but actually no more profound (with regard to consciousness) that the larger light-switches on the wall we use every day, we can easily see that there is no actual consciousness or information of which a computer is really "aware." -- And it doesn't matter how huge the memory-banks of computers become --even 50 or 100 years from now--, there will never be any more actual consciousness in computers than there is in the light-switches on your wall. A trillion electronic switches (transistors) in a computer are no more conscious than one light-switch on the wall... though the illusion of consciousness will undoubtedly become very much more convincing (where it is already quite convincing today).
Oh yes, of course, artificial intelligence (A.I.) will get more and more sophisticated and fast --appearing to have true intelligence-- but the operative word there will always be "artificial" intelligence. Artificial is not actual ...in flowers or intelligence. However, it is truly powerful: --Notable examples of A.I. prowess are in supercomputers, for example: An IBM program called "Watson," competed --and beat-- top human competitors in a knowledge game called "Jeopardy" (using every-day speech); and AI also beat human world champions in highly complex games, such as Chess ...or in "Go" --the world's most complex board game-- where AI defeated Lee Sedol (top level world champion) 4 out of 5 times. - AI also does things like fly airplanes that humans are basically unable to fly reliably. People are also using voice-activated assistants to carry out many types of tasks. "Smart" phones do information-tasks much quicker than people can. --But as for me: At this point, I will be delighted if a computer will simply take verbal dictation and write down everything correctly ...which (despite Watson) doesn't happen at a 100% level yet --but this will undoubtedly come fairly soon.
--But these are still only machines, and they will never possess actual intelligence, which must include awareness. True intelligence must unavoidably include consciousness and self-awareness, which is an attribute of a spirit, which will always transcend the mere machinery (even the neuron "machinery" of the brain). So, an AI computer like "Watson" doesn't actually "know" anything. It only creates digital patterns, and then matches digital patterns, which then produces a "conclusion." But no actual thinking is involved. Machine "learning" pioneer Doug Lenat of Cycorp Inc. in Austin, Texas, says that such an A.I. computer just strings words together and uses linguistic rules and statistical algorithms to arrive at a certain level of statistical likelihood that the informational pattern will correspond to the question being asked. Then when the computer reaches a pre-determined threshold of certainty, it flips another electronic switch to indicate that it has the "answer." --However, the AI program never makes an acual "conclusion" --it is still all a bunch of cleverly arranged switches and logic circuits. --Ones and zeros. --Electronic charges. There is no actual intelligence which is necessarily linked to true conscious awareness. The machine doesn't actually "understand" anything. There is no actual learning or intentional response or answering taking place. --And "emotive" computers --despite all the clever mimicry-- will never be able to experience actual emotions --not even to the level of a dog. --No AI program will ever truly love any person.
10 or 20 years from now, computers will undoubtedly be able to carry on much more sophisticated and smooth "conversations" with people about an ever wider variety of topics, and many computers will be able to perform many kinds of jobs better than humans. For sure, the "Turing Test" will have been easily passed. No doubt, some people will feel love for their robot (AI computer), and they will imagine their robot "loves" them. As a result, some people will undoubtedly insist that those computers have become "conscious" --however, such "consciousness" will still be a fantasy --an ever more effective illusion of "intelligence" which many people will be deceived into or choose to believe. --But this will be an act of self-deception. This is because those computer "conversations" will be no more real interactions with conscious entities, than if we tied a question written on a piece of paper to one end of a string and pulled the string so that the correct answer popped up, which was tied to the other end of the string. The string may appear to have "answered" the question. In the long run, that is all computer intelligence is: Mechanism; not mind. -- Since computers will never have any conscious spirit with self-awareness, there will also be no actual thinking, learning, loving or conversation with a sentient "mind" in computers. -- Though the illusion of self-awareness in computers will be more and more convincing, it will still only be an illusion. A.I. Computers (robots) will never have the consciousness of even a mouse. -- It will certainly happen that A.I. robots will increasingly dominate humanity in certain ways --but that will only be because people will design those robots to act as if they are "thinking" and to so dominate as they will --just as computerized robots already dominate people on the battlefield, in the office and many other places.
However, WARNING! Extreme A.I. Danger: --Despite not being conscious, AI programs are becoming more effective than humans at many tasks: In 2017, the Deep Mind (Google) AlphaGo AI computer beat Ke Jie --world champion in "Go" (the world's most complex game)-- three times straight. At this point, there is no human left in the world, who can beat the best chess-playing computer. --In 2017, an AI computer program passed the National Medical licensing Exam in China ...and it didn't just barely pass: The program beat many people, scoring 456 --nearly 100 points higher than the passing score of 360. --In 2018, a language-processing AI (from Alibaba) outscored many humans in a Stanford University "reading and comprehension" test, after reading and "comprehending" thousands of pages of information, and then scoring 82.44 compared to the humans' average of 82.304 on a set of 100,000 questions. --In 2018, the "Duplex" AI speech simulator (by Google) now sounds very human, including stammers, pauses, breaks, and minor exclamations, so that its speaking seems remarkably natural & human --with the potential to convincingly imitate and impersonate the actual speaking of specific people! As of Nov. 2018, a Chinese news agency has an AI news anchor, working 24 hours a day. It is predicted, that by the year 2020, the AI systems that drive vehicles autonomously on the roads, will drive better than 90% of the people who might drive themselves; as many trucking jobs are likely to be taken over by autonomous systems. --All of these capacities will only continue to improve for the AI programs --which are now even designing better AI machines.
It doesn't matter that AI is not actually "conscious" ...the machines are participating in --and taking over-- many practical human tasks. AI computers use natural language to "understand" and answer questions, make recommendations and perform actions. Machine "learning" is increasingly able to match patterns in everything, such that facial and voice recognition will be ubiquitous, as well as tracking of every behavior people do and everywhere they go. AI now contacts, listens to and evaluates most people in modern countries every day. It is said that AI data on most people is so extensive, that the computers "know" you better than your family members do. --AI computers are now learning how to predict and even control your words and actions! --The horrifying thing, is not that AI systems may become conscious, but that the evil motives of the humans employing the AI systems will weaponize them to control and manipulate people's lives, and to harm their competitors or "enemies." -- AI systems are also being put in autonomous military weapons systems --systems which will know where to find almost everyone on earth.
So, the possibility now exists that AI will become so much "smarter" (though not actually conscious, or intelligent) than humans, that we might conceivably lose control of an AI system that is not merely in one computer in one place, but is distributed among many computers networked around the world ...and therefore its capacities may be "redundant" (in "blockchain" networks) and not able to be controlled by simply destroying two or three super computers. --Hopefully, AI experts will include adequate access (back doors & kill switches) into the programs to always allow humans to get in and maintain control over the AI systems.
But AI is just a tool in machines which now "learn"-- so to speak. But machines only do what humans design them to do: whether the machines are screwdrivers, guided bombs or AI systems. All activity of AI machines depends on the ingenuity of man and what man designs them to do.
But the basic human problem is that man is sinful: "All have sinned, and fall short of the glory of God" (Romans 3:23). So, this will affect AI: People will design AI to control, hurt and profit from others, and gain advantage and power for one's self. -- So, what is the solution?
The Ultimate Solution is found in Christ the God-man, who has the power to set mankind free from selfishness and sin. But the Bible says that "the wages of sin is death" (Rom. 6:23). Christ has this power to free us from the "wages" of eternal death, because he was God in a human body, who died on the cross to pay the death penalty we owe for our sin ...as Scripture says: "Christ died for our sins according to the Scriptures" (1Cor. 15:3).
Christ will fix the whole world some day (read the New Testament), but until then, many things in the world will remain broken, and he's saving one person at a time to fit into that great future: "to all who received him (Christ), to those who believed in his name (his character, promises and actions), he gave the right to become children of God" (John 1:13). -- So, if you truly trust in Christ the God-man, he will help you and change your life for the better.
So, go ahead, enjoy the speed and sophistication of improving computers through the years ...I sure am a computer (and AI) fan. --But much more than that, we should be deeply amazed that we have other conscious persons around us with whom to truly converse, and to share actual intelligence, awareness and true emotions, such as love. No combination of human-assembled bio-chemicals (or of computer processors and software) will inherently produce a true conscious awareness. And no person we meet is merely a biological machine (made up of only chemicals), but these are "persons" who also include a personal mind and spirit, with conscious self-awareness. A soul ...which endures beyond the mechanism (and brain). Every single cell in the body (and brain) is periodically replaced with new cells --but the actual self-aware "person" (their spiritual soul) persists for decades --and millennia, with memories and emotions going way back to childhood. True persons are extremely valuable.
--Therefore, it will never make sense --nor will it ever be just or right-- for A.I. computers or robots to have "rights" similar to human beings. They may look, act and talk more and more like humans --but they will always be only machines; things. --If granting such "rights" to these machines ever happens, that will be a foolish, ridiculous move, as well as a travesty and miscarriage of justice based on an illusion and self-deception connected to clever ingenuity, but not reality. --In fact, dogs and higher animals merit more legal "rights," protections and consideration than any A.I. computers ever will. Computers will always only be machines which are property.
-- And here is the actual source of such true, conscious, intelligent "persons," which are humans: "God created man in his own image, in the image of God he created him: male and female he created them." (Genesis 1:27). God is a being with "personhood," and so is man. This means humans are actual personal beings who are self-aware --having intellect, emotions and a will-- and who have personal relationships and share complex culture with other persons. --The actual consciousness, personhood and moral responsibility of all of us human persons should point us to God, who loved all mankind so much that he sent his one-and-only Son --the God-man, Jesus Christ-- to die on the cross to pay the death-penalty for our sins (Rom. 5:8), so that if anyone truly surrenders to Christ's divine authority and sincerely trusts him, and believes that he died for them personally and then rose from the dead --then those people will be saved (John 3:16). They will then not be responsible to pay the eternal penalty for their sin, but they will instead have eternal life with God/Christ (1 John 5:12) along with all others who have truly trusted and yielded to Christ.
Blessings to you all!
How You Can Know God Personally