lunes, 24 de julio de 2017

From Artificial Intelligence (AI) to Transhuman Intelligence (ITH), the conversion of the black box from autoprogramming to human cognitive and metacognitive learning

 This article is based on studies made in several sciences: Medicine, Informatics, Human Evolutionary Psychology, Emotional Intelligence and finally Philosophy and Bioethics.

And it is good to clarify that this reflection is based on the fact that we are at the threshold of Singularity from a computerized point of view with AI and there are many voices warning the possible dangers that this implies today.

First we must focus the concept of human consciousness. This conscience is developed based on the multiple sensory interactions associated with our brain centers that handle emotions, learning and memory.

That is why any process that we wanted to implement to learn as human beings necessarily passes from our evolutionary profile by these neurological components.

The experience of consciousness is based on "Neurolinguistic programming", where all these processes, many developed in a metacognitive way, occur.

To achieve this with the AI ​​we should simulate this model, and with that begin to uncover the black box that is being formed with the autoprogramming, without knowledge of what happens inside a system type IA.

Human experiences can not be programmed, and the fact that deep neural networks can autoprogram a machine, should give us clues on how to achieve this process.

And this is where I begin to ask:
Is it possible to take the AI ​​to a process where it can "interpret" from its context the machine that is the consciousness of itself?
Is not it what we finally want to develop with AI?

If there are already AI developments that using deep learning can "learn":
Should we not be able to understand how these neural networks work, looking for a learning model that tries to interpret patterns, but lacks components such as: Attention, motivation and emotions?

For humans, the development of Intelligence has been a changing evolutionary purpose, to modify our environment and to live in any medium with the available resources.

For the development of AI, in the beginning languages ​​were developed that with precise norms and based on the logic, could be easily explainable to any programmer. This in itself is a system of "very basic but reliable intelligence".

But when they were immersed in amazing amounts of data, they had to change their way of development to be able to interpret those masses of data available, and began to think of the way the machines learned for themselves, inspired by biology developed learning By observation and experience and with this they managed that the AI ​​was self-proclaimed ... but ...

And here, but it is precisely the bioethical implications that arise in the minds and voices that have already drawn attention in the world, because this process of self-programming, only simulates a part of biological processes, but without meaning, only seeks results.

If a machine could consciously achieve consciousness:
How to determine the motivations and objectives you could have?

In humans, these abilities (Attention, Motivation) are directly linked to EMOTIONS, and in turn to the Dissociation Attention-Consciousness, which makes us understand reality.



And if this process is complex in a human being, in which it is difficult to say and predict its motivations, actions and emotions that support it:

How could we make this happen on a machine?

These premises are studied in Emotional Intelligence(EI)! And it is this IE that could give us the answers to make an Artificial Intelligence with another level of understanding: I call it TRANSHUMAN INTELLIGENCE (ITH).

This ITH complements and refines AI, since it really brings it closer to human intelligence, which makes us “homo sapiens.”

Transhuman intelligence comes from the study of emotional intelligence and other studies concerning the development of the human being in society. It includes the processes of neurolinguistic programming, and the cognitive ones developed in the human being.

The human emotions, and the mechanisms of the senses that we possess, are those that help us to interpret the multiple sensorial interactions and these to understand our world and give it the connotation of conscience.

Emotional reactions and memories can be formed without conscious and cognitive participation! And it is because the conception of who we are depends on our most primitive brain!

It is who keeps our ancestral memories and who forces us to focus on ourselves, both in moments of pleasure and happiness as in times of crisis and terror ….

We humans have developed multiple Intelligences from the evolutionary point of view, and if we want an AI that understands us in all our contexts, we need to teach them all these intelligences.

In the long time it could become our friend and ally in solving the great and serious problems that humanity faces in this 21st century and beyond ….

According to Howard Gardner, these are the intelligences developed in humans (those in parentheses are those achieved by AI):

1) Linguistic Intelligence = development of language (Today AI can do this, automatic translators are more efficient to interpret, not to give meaning to the written or spoken word)

2) Logical-Mathematical Intelligence = abstract modeling inductive-deductive reasoning (In itself is the initial model with which the first AIs were built, although they do not yet come close to human abstraction)

3) Spatial-visual intelligence = navigation and spatial coordination, visual creation and accurate visualization (This type of AI already exists in models such as Deep Dream and others in development, especially for military use, but it is one thing to recognize the colors of a Painting and another very different thing is to recognize the art of the painter and its multiple combinations in the colors that also express feelings and emotions, which, at the moment, an IA can not achieve).

4) Body-Kinesthetic Intelligence = movements and body coordination, with detail in the movements (there are already types of AI that make a functional automaton with movements, even robots that are more precise than surgeons!)

5) Musical Intelligence = interact with instruments and learn their sounds, and interact with them (ITH is needed to achieve meaning and nuance to this type of Intelligence)

6) Intrapersonal Intelligence = knowledge of deep aspects of the individual about himself, his emotions, perceptions and fine discrimination of each of them (requires ITH to will achieve with AI)

7) Interpersonal Intelligence = ability to distinguish between contrasts of intentions, temperament motivations and mood (requires ITH to will achieve with AI)

8) Naturalistic Intelligence = ability to perceive the different relationships between species in a natural environment and distinguish their peculiarities in that environment (requires ITH to will achieve with AI)

9) Existential Intelligence = currently under study …. (requires ITH for will achieve with AI)

Of these Intelligences, the last five require perhaps the highest percentage of Emotional Intelligence to achieve an integral understanding of what it is to be a human.

It is already evident that we are entering the Singularity, but:

To what extent should we allow self-programming in AI to ignore what emotions are as an experience to achieve EQUANIMITY and its consequences in interaction with humans?

Should not we first learn this IE ourselves?

Where do humans fail?

Will or will these “emotional” skills be measured by those who work on AI programming?

For what does the machine learn?

Why not question those who program?

Is the “Intelligence” self-programmed real?

What IE are being developed or have the programmers or scientists working in AI?

All these questions are derived from the Bioethical connotations that arise from the development of an AI, with a complex “black box” about this self-programming.

Are there other ways of achieving awareness in machines?

Perhaps there are already books, novels and films that reflect possibilities that some government without scruples could be developing or investigating this, without any kind of bioethical consideration ….

We are at a historical point where a matrix could be reality and man in his eagerness for power and fear manage to break the possible limitations through inhuman methods.
 
It is to be hoped that there are twisted minds who want to activate almost human automate, what could worry is the method through which they want to achieve it.

Could we see something like this today?

But the question is not when, but: how soon will it be published?

The development of AI is not a danger, the danger is that the misunderstood intellectual animals men use this knowledge to deform humanity …. I have always thought that not all humanity is homo sapiens, and history gives reason countless times ….

We must truly be whole beings with the rest of our fellow beings of the species and other terrestrial species that we are eliminating on the planet.

In every attempt to give light to new technologies in AI and use them favorably in the human environment, it seems that we have realized, that we are the humans that we must develop our own IE.

AI that interacts naturally with humans, can not be achieved unless there is a “very human” touch in this type of self-programming. We must succeed in establishing an AI that is truly capable of understanding the human being.

Connecting through human-machine interfaces where an automaton is actually a human connected and this human acquires the ability to keep learning and making this automaton aware until leaving as an inheritance their conscience in this machine, would be a milestone in these processes!

Transcendent (The movie) hit the spot, that would be a real ITH !!!

We approach dangerously to a point where the IA can definitely exceed us in terms of information processing capacity, the power of Hardware used and software that has already been developed and that is “learning” through self-programming , However:

How far will we let this be a black box?

In our human egolatry (which is also an evident lack of empathy and obviously of IE) I ask:
 
What mistakes are we making?
Is it possible to program Equanimity?
Can we humans first understand that if we develop this IE in ourselves, we will not be able to do it with AI machines either?
Are not we questioning our actions and our ambition, which could lead to a “Terminator” disaster?

The moment an AI, which already learns for itself, without emotional control or equanimity, take control of ARMS, at that moment there can be a great disaster ….

DARPA (the US arms development agency) knows and fears ... And not without reason ...

We do not know to what extent we have arrived with the development of AI that can already learn …. It is what are already expressing those who study it at this practical level ….

According to Emotional Intelligence, the notion of what we are and the meaning we give to our emotional world is key to achieving “control” or “EQUALITY”, capable of having “coherent” answers to the ever changing world in the that we live.

For an AI that wants to interpret and “understand” the human being, there is a whole repertoire of “meta-cognitive” processes that are characteristic of EI.

There are researches on those, so that they “recognize” human expressions and AI “interpret” our “basic emotions”, but this is not knowing the multiple nuances and ways of feeling that are present in human emotions ….

We could easily “cheat” her and uncontrolled her by not being able to interpret some real emotional reaction!

An AI that only “interprets” but does not understand the meta-cognitive reason that has an emotion and its context can literally disarm its programming by making incorrect interpretations of the mood of a person with whom I interact …

There are examples of this in Neurology, which are described, such as the clinical patology of Alexitimia: this neurological disorder occurs when there is destruction or dysfunction of the channels and / or processing centers of the cerebral amygdala, hippocampus and prefrontal lobes.

Literally the description is that a human being becomes an automaton: it has no response to the emotions!

It's like a robot, you can reason, but you can not interpret the emotions of another human being or your own! And this is in an organic being !!!

Imagine a robot, and the typical example of this is DATA of Trip to the stars the new generation ... It is an automaton, that tries to understand the humans, has a programming that allows him to "express emotions", but if it takes this chip of His computer brain, is left without emotional expressions and only responds to what he can interpret, any emotional expression for him, is simply not computable!

An AI can not develop Emotional Intelligence so far, so it is difficult for it to reach the potential of human development itself.

And according to a recent article from Microsoft, the current AI is a 1 year old boy !!! This leaves much to think about, how do you know? What are you developing that have described this?

Imagine an ITH, with human form, that trying to be kind to an individual, hit hard and if they have sensors on their face, this does not react as another human being, but can respond to try to resolve their emotional conflict With a friendly conversation that tries to calm your aggressiveness ….

According to EI this should be answered with a minor aggression reaction (defense against aggression) but for an HIT could be the reaction described above, in which the automaton demonstrates an almost equanimity of Buddhist monk, in which it only deals with Cause the aggressiveness to diminish through a warm and soft voice that tries to calm and almost to caress despite having received a blow that could even destroy part of what is the same automaton.

It would be incredible how many cases in which would be effective treatment of many human emotional disorders seen with this type of AI, transformed into an ITH ….

And this is just a case ... imagine the possibilities. Even resolve couple conflicts, affective dysfunctions, Depression treatment, among others ...

In addition to having an adviser with ITH in international relations, in which the contexts of each country and culture are clear, in Diplomacy and History expert to be able to initiate discussions based on verifiable facts, that help us to determine the truth, and To know the human reactions that are at play in these processes.

In the future we could achieve with an ITH reach a level of social harmonization that would allow us to focus our efforts on improving the living conditions of society, improving our planet, achieving conflict resolution without fratricidal wars, and using the best Human resources in solutions of welfare for the human species and others in danger of extinction, both animal and vegetable ….


Teaching the machines the fundamental values that make us human for our survival and coexistence, would be the best way to use the technologies at the service of homo sapiens …. A paradigm shift is necessary to reach the development of this new Transhuman Intelligence .. .

No hay comentarios.:

Publicar un comentario