Artificial intelligence is about replicating the adaptive reasoning process that humans or higher animals use. For that, we need to understand what intelligence is, and how it is used.
We hear a lot about artificial intelligence (AI) these days, and about how it will impact our lives, both in work and play. However, AI has been touted as an impending societal game changer for decades, but very few of what has been predicted has come true, at least not on the visually grand scale envisioned.
Most changes from AI have been behind the scenes, as any large scale public changes, such as driverless cars, typically need legislation, regulations and policies in place, and big decisions in those areas take a lot of time. Problems with the behind the scenes ones have taken time to manifest, yet the magnitude of some of them has created headlines. These have included mass advertising targeting, identity leaks, spam and fake news bots.
Progress in AI development depends upon how much we understand of the mechanics of intelligence, particularly in regards to the brain. However, since a lot of the mechanics of the brain is still unknown, it is difficult to predict how stable AI will be when it encounters the variety of real world situations.
Intelligence is often associated with thinking, but it is more related to cognition and perception, which is much more complex than linear thinking, as used in traditional computer programming.
We may think we taste with our tongue, but the vast range of what we think we are tasting is because we are internally modifying the basic things we sense with our tongues with the aromas we smell at the same time. We may hear continuously variable analog sounds, but they are converted to digital in the ear. We may see a wide range of colours, but they are converted to a high definiton grey-scale matrix, and a low resolution three-colour matrix in the eyes. That is, we recognise some basic properties through our senses, but modify that to perceive something within that as meaningful to us.
Evolution has brought us to perceive what we want, or don't, in a sea of noise. Thus a spider can ignore the buffeting of its web in the wind, yet detect the faint struggle of prey. An animal can detect the faint smell of a lion in the midst of the stench of its own heard. A person can hear their name being called in noisy traffic.
While a standard IQ test may have a variety of questions, and a few related to picking the odd one out in a set of patterns with a variance in shape and position, those for higher IQs typically include lots with patterns with a half dozen variances.
Natural intelligence is about being able to perceive exceptions in amongst lots of noise. And this is where AI comes in. Typical software is about how to process the noise, and ignore the exceptions as errors. AI wakes up upon the exceptions, and so is better able to handle the vagaries of the real world.
In the real world, we often do not know what will happen, so intelligent systems need to be able to adapt to evolve.
Having a system that can detect patterns, or their lack, in its inputs, is the starting point of AI systems. It may be able to pick the odd one out of a set of a few simple variances. Over time, an AI system may build up a library of such complex pattern detection sets.
Where AI really comes into its own is when any one of those simple variances can be replaced by a complex one. Since that can be done exceedingly quickly, AI can evolve exponentially by a rapid series of such swaps as it adapts to its inputs. For example, a AI system using a 6-variance pattern can change to one of 36 variances, just by swapping each variance detector with a 6-variance one in one iteration. That makes for a very adaptable system, and one that leaves programs that modify their own code logic in the dust.
While conventional computer subsystems have some basic computational capability, AI systems can really benefit from an AI approach to its subsystems as well.
Traditional inputs for computers have been keyboards, mouse and more recently, touch and voice. However, applying some AI can make input interfacing much more adaptable.
A typical automated information kiosk will have large buttons drawn on the screen, perhaps with multiple languages on the buttons, or somewhere to select an alternate one. An AI one may have an audio input, so that it listens to who is around. If it hears a particular language as closer than the others, it may change the language shown on the buttons to that.
One area where people have difficulties is when they are talking in their language with someone whose native language is not the same. While people may learn the words of a new language, they often use their native grammar, which will likely have different word orders for things like noun-adjective and verb-adverb combinations.
People usually find such grammar changes make communication very difficult. However, an AI system could be seeded with a variety of grammars, and upon detecting which one is spoken, use it to interpret the meaning. Of course, it would use the correct grammar in replies.
Such a kiosk may have a camera, by which it may detect whether a person may have a disability, like Parkinsons, where movement may be varying, and so detect the range of motion that person has, and so enlarge the buttons and adjust the hit detection timing to compensate.
Depending upon the complexity of its AI data stream requirements, the kiosk may need a direct connection to a database, besides the connection to relay input to the program using the kiosk as its front end.
Storage has some smarts, but mainly to do with managing its own mechanical operations and data flows. Databases usually manage the structures of data, and recognise commands that specify what to do with its data.
Modern databases have named tables of regular data, and may have named procedures to process it. They are usually responding to explicit commands from the requesting program. AI data streams and associated pattern-cognition can be quite complex and demanding, so the main program could offload some smarts to an AI in the database. That could be in the form of named data streams, and preemptive loading of those that are mostly used. In a full AI-based setup, the AI support database will be used by several subsystems, so it is better to include the AI database as an integrated part of the storage.
Being able to adapt to a variety of inputs and situations usually needs an accompanying means of changing the environment. To do anything sophisticated requires some form of robotics.
How adaptable the robotics have to be depends upon how many types of alternate actions the AI needs to implement. Fortunately, we can again look to biological evolution for the lead here. Over many generations, animals sometimes developed very specific body parts for the particular circumstances they most commonly found themselves in. Crabs developed large claws for fighting over food, or defending themselves from attack. Some species developed feathers for flying, and others body colours for camoflague. These are adaptations that have proved to be enduring.
However, some adaptations, like the large claws for crabs, worked well for some functions, but not well for others. They then stay away from circumstances that they cannot handle well. Other species, like primates, evolved to be more generic, with the opposable thumb opening up a whole lot of flexibility in how they dealt with their environment.
While many species can use tools because they have enough intelligence for basic reasoning, walking upright and having adaptable hands gives opportunities to use a lot of tools, especially since hands provide the opportunity to make more sophisticated tools. So we come to the human form, being perhaps the most adaptable on the planet, which is why it has been able to reshape the planetary environment to such a huge degree.
If what we put AI into needs to be able to change from a soldier carrying a weapon to a medic applying basic first aid, the robotics need to handle changing tools, so the human form would seem to be the ideal, especially as we already have a huge range of suitable tools available right now.
The problem is that of acceptance, as people tend to feel threatened by something that could possibly be better at some activities than them. Perhaps the real problem is that such robots don't go through the angst that people go though when confronted with change, so will be more likely to adapt to new challenges faster than people.
While artificial intelligence can simulate a lot of human thinking processes, and at speeds far exceeding those of people, they cannot tap into the highest realms of human consciousness.
Human consciousness is more than the sum of our brain synapses, but a trans-biological awareness that can tap into knowledge realms that any mechanism tied to purely mineral-based physical objects cannot. That means computers are essentially hamstrung by their basic structure that prevents them from any sort of higher consciousness.
Computers cannot be self aware because they do not, and can never have, a higher self. While animals cannot reach the levels of consciousness of human beings, they can still tap into a trans-body awareness that minerals, and anything based upon them, cannot. It is only because of our limited concept of ourselves and our capabilities that we think that computers can somehow become better than us.
We have made great strides in building devices to augment our bodies, and even to replace some organs with fairly rudimentary machines. However, while those replacements may last longer than their biological counterparts, they cannot work at the same level of synergy with their enclosing body that biological parts can. They will get better, but their progress can only be asymptotic to biological capabilities, let alone human.
The standard view of the brain is that of a fancy adaptable computer that has dedicated areas to tend to our various bodily senses, functions and higher needs. We have discovered that many of our bodily sensory systems actually do an analog to digital conversion to encode what we have sensed into pulses transmitted along our nerves, further giving the impression that our bodies are one gigantic distributed computer governed by a core housed in our brains.
This seems to further reinforce our idea that consciousness is just the sum of all that biological activity. But our biology is more than a self-contained computing apparatus, but a doorway into other realms of consciousness, each with their own mechanics, but at an ever more subtle level. It is better to think of each person like a computer on a network, but connected to a cloud of consciousness.
If we have a higher connected consciousness, that begs the question of why we seem to be limited to seeing ourselves as separate and fairly unconnected to each other.
That is all for us to focus on understanding this physical world without distraction from the expanded consciousness of our higher selves. We are given free will to decide our individual destiny, but with a hope that with that free will our personalities will choose to seek out our higher consciousness and use our personalities as its instrument in the world in service of our higher good.
In all of our study of the physical world, we have been learning to become creators, progressively becoming more like gods. That is deliberate, but not with the intention of us gaining power over others, but power over our personalities, just because that enables us to bring the divinity of our higher selves into the physical world.
The lifetimes of our personalities are limited, with the two-fold intent of limiting the amount of control any one personality can achieve over others, and resetting the grief levels that we would otherwise accumulate from all the suffering we have created, which would cripple us emotionally to the point of self-destruction.
By exploring artificial intelligence, we are delving into the processes that govern our consciousness. It is a merging of our understanding of computing and psychology. However, its study can also show that we are more than the models that we program these advanced computers with. We can transcend the limited consciousness of our personalities and reach into the larger consciousness. One that is not tied to our selfishness, but the well-being of all.