Some would say that AI – artificial intelligence - already exists. Many others believe that it is only a matter of time before it is achieved.
Don't be fooled.
Real AI will not be achieved. Ever.
Some very impressive computers/robots are already on show. And there is no doubt that far, far more impressive things are still to come; perhaps sooner than later. These things will give overwhelmingly compelling impressions that they are genuinely intelligent entities.
Many people, if not almost everyone, will be persuaded on the basis of this seemingly incontrovertible evidence in front of their eyes that it is true: real artificial intelligence will have been achieved. I don't doubt that I too will most likely be strongly tempted to agree.
However, don't believe it. No matter how convincing it may seem to be, apparently real artificial intelligence will essentially be just a very, very clever illusion or trick.
How can I be certain that that is so?
Think about it. A hole is to be dug. A person takes a spade and makes the hole. Who or what has dug the hole? While it is true that it is the spade that contacts the soil, it is the person that has dug the hole. The spade is just an unconscious tool used to carry out the person's intentions. If a larger hole was needed a back-hoe may have been used but it would still have been the person who dug the hole with the back-hoe being the tool.
A person uses an abacus to work out a maths calculation. Who or what does the calculating – the person or the abacus? Even though the beads on the abacus display the right answer in no sense does the abacus do the calculating. It is just an instrument used by the human operator.
Is it any different when more complex machines such as calculators are used? Despite their name, no, these machines do not "calculate". The machines themselves do not think. Just as the beads on the abacus are manipulated by a human operator, this time the manipulating is done electronically but still in accord with a programme written by human beings.
Computers are much more complex still but the same principle remains. In response to input things happen in a computer in accordance with the programme and output occurs. It appears that computation has been carried out by the machine but it is appearance only. The computation has been carried out by people – the operator providing the input which works with the human-designed programme.
If a computer output shows a wrong result we know the mistake was made by one or more of the following - the human operator, the human programmer, or the human hardware manufacturer. We do not blame the mistake on the computer's "thought processes". (Curiously though when a computer appears to get it "right" such as when for the first time a computer, Deep Blue, "defeated" world chess champion Gary Kasparov in 1997 people are much readier to credit the computer with the ability to think.)
No, just like a spade or any other machine, a computer is nothing more than a lifeless, non-conscious, non-intelligent, non-thinking lump of matter (albeit intricately assembled) that is useful as a tool for human beings. And this will be true of any future machines/computers that humans create.
Yet it is obvious that human beings are intelligent, thinking entities. (If we don't believe that is so . . . well, what can one say?) So, it will be objected, if humans are intelligent why should it not be possible for them to use that intelligence to create other intelligent entities? After all, if all life, including intelligent human life, has just happened by itself to evolve out of pond scum, then why wouldn't we be able to deliberately bring about other intelligent life?
The key question now is, what are the criteria for something to be identified as being intelligent? One necessary requirement is that it must have agency, i.e. it must have free will.
For something to be a thinking thing it must be able to initiate active power to make intentional, voluntary decisions. For example, this is what a human being does when they purposely or intentionally raise their hand to make a bid at an auction or choose to refrain from raising their hand and not bid. Thus we can identify them as a free agent.
If there is no free will though, then everything that happens is determined, i.e. everything happens simply as a consequence of the laws of physics acting upon matter. If the "thoughts" that arise in human minds are just the end result of a long chain of unbidden, uncontrolled mindless physical processes then "thoughts" are nothing more than so much internal noise. An agent however is in control of their thoughts and is not just a passive object through which the chain of states of affairs passes.
But no such agent can exist if a physicalist account of the universe is accepted, i.e. physical matter is the only substance in the universe. Matter alone is passive and inert and cannot generate causal powers. It is only if an entity has a non-physical substance, something that is to some extent outside the mechanical unfolding of the universe, that it is possible for there to be free will and intelligence.
Can human beings create such a dual substance entity which has originating causal powers? For those who believe in the possibility of real artificial intelligence, many hope that by generating a sufficient degree of complexity in a system that will somehow bring intelligence into being. However that is just a massive leap of faith. We have no idea how a physical substance could have evolved to generate a non-physical substance that has the requisite ability to display free agency. Neither do we have any idea how we could create such a non-physical substance.
Machines are fully causally determined. They do not think. They cannot be right or wrong in what they produce. They have no goals.
If people are machines then people do not think. Yet people indisputably do think. Therefore people are not machines. People are free agents and this provides evidence that they are made up of two substances – one physical and one non-physical.
When humans use a machine such as a computer it is the human who is carrying out a procedure, with the machine, no matter how complex it may be, being a mere tool that makes it easier for the human to reach their goal.
We have no reason to believe that humans will ever be able to create the non-physical substance that is a necessary requirement for something to be a free agent and an intelligent entity.
Don't be fooled. No matter how incredibly impressive computers/robots will become, they will never be intelligent. They will always remain lifeless, non-conscious, non-intelligent, non-thinking lumps of matter – like a very fancy spade.
It is always and only the humans – they who make and use the computers/robots - who are the intelligent entities.