The right picture of AI (deep learning) & the right business model of big data predictive analytics
Table of contents
Consequences of a false idea of AI & sexual intelligence. 2
The persistence of the false idea: the Darwinian connection. 2
False and right business models: an example in credit management. 3
Concluding with a paradox to show AI is IU.. 4
Executive summary
Artificial intelligence is in essence based on a false picture how humans can know reality leading to the thought that machines can be intelligent self-conscious ‘beings’.
This picture is described. This bewitching picture has consequences on the level of government (granting grants for research in AI).
The irrational holding on to this picture is ‘caused’ by another, Darwinian, picture that pictures humans as being (chemical) machines. Once the picture that humans are machines is destroyed the picture that might control you in thinking that machines can be intelligent self-conscious ‘beings’. Reasoned is that there are no logical arguments to destroy the Darwinian picture for pictures, images, are the input for the logical machinery and proceed logic.
However in business we do need to make very critical business decisions in which we must choose between the thought that AI can take over business, or at least part of it, and the idea that humans are in the lead in which AI is another, very handy, tool. Repetitive routine jobs can certainly be automated with AI, but when human ‘psychology’ is involved better to have a business model in which humans, whether or not as parts of interdisciplinary teams, take the lead.
An example of in credit management and the collection process is given. And the moral question is raised whether you may use your clients as input for the ‘learning’ process of AI machines.
This shows that very concrete specific business decisions are dependent on your stance in seemingly very abstract ‘philosophical’ subjects. ‘Keep it real or no deal!’.
The A in AI & the I in AI
(6-9-2017). This article is a short reaction on a very concise article, by Willem Peter De Ridder, that describes both the history of AI and the current state of affairs: https://www.dhirubhai.net/pulse/man-machine-who-decides-part-1-series-perspectives-de-ridder.
One and only one sentence however should be skipped or rephrased: 'For the time being, people are better than machines when it comes to determining goals, dreaming about the future and taking morals into consideration'. 'For the time being' should be written as 'Till the end of times...'
Machines will never become self-aware and knowing is not identical to gathering data and knitting it together. Only in a shallow self-reflecting on your own 'process' of knowing you will construct this ‘process’ retroactively as combining data in an intelligent way. That reconstruction shows the artificiality of artificial intelligence (as the general concept that includes 'deep' learning as well). Knowing is not a process in time, but is immediate (in zero time). (Do not picture, unrealistically imagine, reality is otherwise: though shalt not make images of things created. In this article I will not elaborate on proving immediate knowing of reality).
AI imitates (and in the case of AI you even do not need intelligence to imitate!). It would be better to call AI 'Unintelligent artificial intelligence' (UAI) so you can always be triggered to become aware of the fact that machine are stupid (even if we cannot predict their calculations). UAI is less misleading as a term.
So if the ‘A’ means ‘imitating’ and the ‘I’ actually means ‘unintelligent’ ‘AI’ stands for ‘IU’, that is ‘imitating unintelligently’.
Consequences of a false idea of AI & sexual intelligence
You might think that I am over exaggerating, but large sums of budgets are given to AI projects on the basis that machines could be thinking in the future. (The human brain project raised about one billion euro`s). I am not saying no scientific discoveries could be made in such a project, on the contrary, but do not give away tax money for the wrong reasons.
If I would be in some kind of commission to assign grants for AI I would ask the applicants whether they believe in sexual intelligence. If the answer would be ‘yes’. I would ask whether machines can have sexual intelligence. If the answer would be ‘yes’ I still would not dismiss the applicant, for his research could still be useful for scientific discoveries. I would dismiss the applicant if he would say ‘yes’ to the next question: ‘would you have sexual intercourse with a machine’. I would dismiss the applicant even when he would hesitate for a few moments and say ‘no’ for I expect an immediate ‘no’ (even if the machine would look like your favourite movie star).
The persistence of the false idea: the Darwinian connection
The false idea of AI leaves us with a false enchanting picture in which we humans are ‘epistemological autistic’ meaning we are blindly imprisoned in ourselves, have no access to the world and can only have access once we gather data, that we combine intelligently to form a worldview that somehow fits on the world. That ‘fitting’ makes us see the world. The worldview imitates the world.
Of course if you think about it you need intelligence in the first place to know which data is useful and that means you immediately see the world. If you compare data with flowers you must see the world before you can see the flowers you must pick to combine them intelligently in a bouquet that imitates the world. That is ridiculous (for the world is more than a bouquet of flowers and once you see the world it is not necessary to pick data out of it).
This false picture of human awareness (or knowing) in which blind data is the input for empty intelligence and that the combination makes you see the world, in which we are cut of the world and then try to bridge the gap is a true picture on how machines calculate. The false picture persists when you believe humans are machines. Chemical machines, or as I like to call them talking pieces of medium rare steaks. Chemical machines that have been the result of years of evolution. Self-consciousness is, you are, just a epipheneoma meaning self-consciousness is an illusionary effect of the chemical machine, rather than a cause of all the (brain)movements of your body. Your body is not a chemical machine and you are not a machine.
The argument runs as follows:
1 AI exists when men are machines/if men are machines (intelligent self-conscious) AI exist
2 Men are not machines
Ergo: AI does not exist
This however is a false syllogism having the same structure as:
1 If the streets are wet if you have thrown water on the streets
2 You have not thrown water on the streets
Ergo: the streets are not wet
(1 If A then B
2 Not-A
Ergo: Not-B).
This is plainly false for it might have been raining. You can reason that you cannot, so to speak, ‘reason a false idea out of someone’ analogous to the Dutch expression of beating an idea out of someone. To understand what AI is is to understand what human intelligence, as an appearance of bodily self-conscious for bodily self-conscious, is. Then you can immediately see, know, that men are not machines and that machines cannot be men. (Think about the song of Jo Jackson ‘I wonder who the real men are…’).
False and right business models: an example in credit management
Big data is all about data, flowers, in different data sources, flower fields. Big data is the field of AI. AI certainly has a role to fulfil in many fields. The question is where you make the decision between applying AI in the sense that you fully automate a process (including searching the right data sources), thus without human interference, and working together with a machine by forming clever hypothesis that enables you to search the right data sources to get a clearer picture of ‘what is really going on’. Do you trust on your subject-matter expertise (and that of your colleagues) or do you fully trust a machine? I ‘believe’ in team work with your colleagues and machines when it concerns processes where other humans, your self-conscious clients, are involved. When psychology is involved a machines should not be your manager. (Self-driving cars may at times manage your driving). Make your job more sexy be using machines rather than be used by machines.
Let’s take a look at the collection process. The first collection, the first reminder might be fully automated, but once an invoice is still not payed, due to a specific change in the situation of a client (especially in B2B) proceeding with automation can lead to Kafka’s ‘process’. You might object and say the machine is a deep leaner and will adapt to the situation of the client. Well that might be so over time, but a solution is needed ‘Now!’, asap. It seems you will use this client to let your machine learn and use this client, this human, as an instrument, a means, for the benefit of other clients.
I can picture a future in which you both have machines in order to test hypothesis (of why certain clients do not pay in time or do pay in time) and once the hypothesis are tested that information is fed to the deep learning AI machines. In the collection process that not the case yet for too much psychology is involved. (That is the reason why sales is an ‘eternal’ profession for these AI machines should be sold. Can you imagine machines selling themselves?).
For a further elaboration on big data and credit management: https://www.dhirubhai.net/pulse/big-data-goliath-analytics-stone-david-humanity-use-case-roolvink.
Concluding with a paradox to show AI is IU
If you could understand your intelligence completely you would be so stupid that you could not even begin to understand your intelligence. The fact that we cannot fully grasp our intelligence shows we are intelligent. AI deep learning machines are built on the idea that they understand their intelligence.
The article I mentioned in the beginning of this article raises the question ‘Man or machine?’. The first answer is ‘Man’. The second answer is ‘Man and machine’.
Kind regards,
James
Title picture 'Where man and machine should never meet'. I could not find the person who created this picture.