Artificial intelligence is on everyone's lips. What is it about these systems? Are they really intelligent? Are they useful? Are they dangerous?
I am a mathematician and developed such systems some time ago. And since I have been researching in psychology for the last few years, I speak from the perspective of a double expert.
A tool is neither good nor bad. It's how you use it that makes it so. You can use a hammer to drive a nail into a wall – or injure a person. The good and the bad of artificial intelligence result from its use. But first we have to deal with the ugly.
The ugly thing about artificial intelligence is its name. These systems are NOT intelligent. They function, which is the opposite of intelligence.
Such a system comprises three components:
1) fact and rule knowledge
2) mathematical methods for data continuation (such as interpolation)
3) methods to augment fact and rule knowledge
These components are combined in a program that is executed with high computing power. At its core, such a system works like a weather prediction program. There is no artificial intelligence. Just like any tool, such a system also harnesses the natural intelligence of those who created it and those who contributed their know-how in the form of knowledge and algorithms. A bicycle also harnesses the intelligence of those who designed and built it. Every tool does that.
The alleged AI systems impress most people for three reasons: first, by their results.
One example of how easy it is to impress people with results is the ELIZA program, which was developed by Josef Weizenbaum at MIT in the mid-1960s. It was a simple computer program for processing natural language. ELIZA simulated conversations using a simple technique that gave users the illusion that their input was being understood by the program. Weizenbaum wrote that “many early users were convinced of ELIZA's intelligence and understanding, despite his insistence to the contrary.”
60 years later, computer technology has come a long way. Today's programs are much more complex, use fact and rule knowledge from vast databases together with sophisticated data continuation algorithms, and they have many times the computing power in the background. Therefore, their results are much more impressive than they were back then.
Fractals are a vivid example of how results can impress. Do an image search on the internet for the keyword ‘fractal.’ You will see beautiful pictures. But there were no creative or intelligent artists at work. Simple computer programs generated the images through countless repetitions of simple calculation rules.
The second reason these systems impress so many people is a lack of understanding of the term ‘knowledge.’
Most people confuse knowledge with truth. An archaic synonym for ‘know’ is wit that has the root *weid-, which means ‘to see.’ Taken together, to know means to see – with the senses or mental. What someone knows is, therefore, nothing other than a perspective.
What someone says or writes, what is in a book, what is in a database, and also what is taught in schools and universities is not the truth, but a perspective. It can be more or less close to the truth.
The misunderstanding of the term ‘knowledge’ arises because science as the collector and administrator of knowledge today has the same status as the church did hundreds of years ago, namely as the supposed guardian of truth and a social authority. The alleged AI systems as products of scientific research inherit this illusory prestige.
The third reason these systems impress so many people is a lack of understanding of the term ‘intelligence.’
Most people confuse intelligence with education. Education is acquired. It is nothing more than programming. Intelligence is natural and has nothing to do with education. We see this in children, who are the most intelligent people on the planet ... before they go to school.
Intelligence is also often confused with the ability to learn. The ability to learn arises from programmability. Every living being is programmable and therefore capable of learning: every amoeba, every plant, every animal, every human; only humans are also intelligent. I explain the difference between humans and animals in my article “This is the Difference Between Humans and Animals” and in my book “Consciousness – Its Nature, Purpose, and How to Use It.” You can find out more about intelligence in my article “How We Lose Our Innate Intelligence – and Can Regain It.”
The alleged AI systems are not intelligent; they are educated and capable of learning – because they are programmed and programmable.
The most appropriate name for such a system is ‘expert system.’
An expert is someone who has experience. Experience leads to knowledge/perspectives in the form of facts and rules. Feeding them into a computer program yields an expert system.
Human experts know that their knowledge comprises only perspectives, is constantly being changed by further experiences, and is therefore never definitive. They are constantly learning. Although computer systems can also learn, their learning follows the methods that have been programmed. Because of our (natural) intelligence, we humans can also learn without a method. Therefore, our learning possibilities are infinitely greater than those of computers.
Now we come to the good and the bad of expert systems.
One good thing or benefit is the preservation of expert knowledge. An illustrative example is a project in which I was involved. A ski company contacted us with the request to develop an expert system for preparing skis for ski races. One of their employees had been doing this for decades and had become an expert. They wanted to preserve his know-how before he retired.
Preserving doesn't just make sense before retirement. The competition could poach an expert – or the expert falls ill – or even dies.
Another benefit is the multiplication of the usability of expert knowledge. Suppose there is only one expert in the world for dealing with a specific type of accident. If his know-how is preserved in an expert system, one can apply it as often as required.
A danger of these systems is that we could rely on them too much and neglect the training of human experts. When an apprentice learns from a master, the master’s knowhow combines with the apprentice’s intelligence and creativity. The apprentice then becomes not just a human expert system as a copy of the master; he can advance into dimensions that exceed those of his teacher and that computer expert systems cannot reach.
Another major danger of these systems is that the programmers can color them ideologically at will.
Last but not least, these systems are a very serious psychological danger as long as they are referred to as ‘artificial intelligences.’ Up to now, most people have probably realized that although machines are superior to us in terms of qualities such as strength, speed, and endurance, we still have a unique intellectual position. By persuading people that machines are now also intelligent, machines are not only put on our level, but even above us, because machines are more powerful in what they can do. The term ‘artificial intelligence’ alone makes people believe they are inferior to these systems and thus small.
This is fatal social programming. Humans are limitless. Computers are limited. Computer systems show us which tasks have nothing to do with intelligence and creativity, but are merely routine. In fact, these systems show us that we act robotically most of the time, in that 99% of our actions and even thinking arise from social and other programming. I explain this in my article “Why the question ‘What am I?’ is important and magical.”
Artificial intelligence is an intelligence test for us. It will divide humanity.
· Some people recognize these systems for what they are: tools that can help us grow – as productivity tools as well as tools for reflection. These people can recognize their own programmed nature. If they do, they can step out of their robotic existence and become true Homo sapiens.
· The others descend by placing themselves on the level of machines. In doing so, they reinforce and consolidate their robotic existence and finally become Homo automaticus.
Further reading:
Article “This is the Difference Between Humans and Animals”
Book “Consciousness – Its Nature, Purpose, and How to Use It”
Article “How We Lose Our Innate Intelligence – and Can Regain It”
Article “Why the question ‘What am I?’ is important and magical”