Will AI Simulate the Subconscious Mind?

Your sub­con­scious mind is the actu­al play­er, which is con­trol­ling you from behind the scenes. It  is the part of our mind that makes deci­sions with­out our hav­ing to active­ly think about them.

Arti­fi­cial Intelligence(AI), giv­en prop­er time­line of evo­lu­tion and the right algo­rithms, might sim­u­late a strong enough ver­sion of the sub­con­scious mind.

Conscious, Subconscious and Unconscious Mind

The con­scious mind involves all of the things that you are cur­rent­ly aware of and think­ing about. This aspect of mind is some­what akin to short-term mem­o­ry and is lim­it­ed in terms of capac­i­ty. For exam­ple, our aware­ness of our­selves and the world around us are part of our consciousness.

Unlike the con­scious mind, the sub­con­scious is that part of the con­scious­ness that is not cur­rent­ly in focal aware­ness. An exam­ple of the sub­con­scious activ­i­ty is the part of the mind that cre­ates dreams. Like dreams, sub­con­scious activ­i­ty occurs in the mind with­out con­scious per­cep­tion, or with only slight per­cep­tion, on the part of the indi­vid­ual.

The uncon­scious mind involves the process­es in the mind that occur auto­mat­i­cal­ly and are not avail­able to intro­spec­tion, and include thought process­es, mem­o­ry, affect, and motivation.

The uncon­scious mind dwells out­side of our con­scious aware­ness. It is a reser­voir of feel­ings, thoughts, urges, and mem­o­ries. Cog­ni­tive psy­chol­o­gists say that the uncon­scious mind con­tains data that are unac­cept­able or unpleas­ant, such as feel­ings of pain, anx­i­ety, or conflict.

Sig­mund Freud (May 6, 1856 — Sep­tem­ber 23, 1939), Aus­tri­an neu­rol­o­gist and the founder of psy­cho­analy­sis, believed that the uncon­scious mind stores all the thoughts, mem­o­ries, and feel­ings that are dis­turb­ing or trau­mat­ic. Freud pop­u­lar­ized the term ‘uncon­scious’ when he devel­oped psy­cho­analy­sis. Psy­cho­analy­sis is a form of ther­a­py devel­oped by Freud to treat emo­tion­al problems.

Related Readings:

Freud empha­sized the impor­tance of the uncon­scious mind, and a pri­ma­ry assump­tion of Freudi­an the­o­ry is that the uncon­scious mind gov­erns behav­ior to a greater degree than peo­ple sus­pect. Indeed, the goal of psy­cho­analy­sis is to make the uncon­scious conscious.

Although Freud gave more empha­sis to describe the uncon­scious mind, mod­ern notions of the sub­con­scious were also invent­ed by him as part of his now-dis­cred­it­ed the­o­ry of psy­cho­analy­sis. Accord­ing to Freud, 90% of the human mind is the sub­con­scious, full of invol­un­tary thoughts and behaviors.

Mind Researches since 20th Century to the Present [in Brief]

Famous ana­lyt­i­cal psy­chol­o­gist and psy­chi­a­trist Carl Jung, who was an ear­ly 20th cen­tu­ry psy­chother­a­pist and psy­chi­a­trist, cre­at­ed the field of ana­lyt­i­cal psy­chol­o­gy. He the­o­rized that when we dream, we tap into a “col­lec­tive uncon­scious” and devel­oped arche­types for the sym­bols that com­mon­ly occur in dreams.

Relat­ed Arti­cleIs our col­lec­tive con­scious­ness dif­fer­ent from the con­scious­ness of each individual?

In 1953, a young Ph.D. stu­dent Eugene Aserin­sky in U.S. dis­cov­ered rapid eye move­ment (REM) sleep. Dur­ing REM sleep the brain is extreme­ly active and pro­duces dream.

In the 1990s, neu­ro­sci­en­tist Melvyn Goodale began to study peo­ple with a con­di­tion called visu­al form agnosia. Such indi­vid­u­als can­not con­scious­ly see the shape or ori­en­ta­tion of objects, yet act as though they can. 

Sci­en­tists have also start­ed to test ways of detect­ing con­scious­ness with­out the need to give peo­ple ver­bal instruc­tions. In a series of stud­ies that began in 2013, neu­ro­sci­en­tist Mar­cel­lo Mas­si­mi­ni at the Uni­ver­si­ty of Milan and his col­leagues have used tran­scra­nial mag­net­ic stim­u­la­tion (TMS) to cre­ate elec­tri­cal ‘echoes’ in the brain that can be record­ed using EEG.

In a 2017 sleep study, researchers roused peo­ple through­out the night while mon­i­tor­ing them with EEG4. Around 30% of the time, par­tic­i­pants who were jolt­ed from sleep report­ed not expe­ri­enc­ing any­thing just before they woke up. The study showed that those peo­ple with­out con­scious expe­ri­ences dur­ing sleep had lots of low-fre­quen­cy activ­i­ty in the pos­te­ri­or-cor­ti­cal region of their brains before waking.

In a 2019 study that col­lect­ed fMRI data from 159 peo­ple, researchers found that, com­pared with peo­ple in min­i­mal­ly con­scious states and those under anes­the­sia, the brains of healthy indi­vid­u­als had more com­plex pat­terns of coor­di­nat­ed sig­nal­ing that also changed con­stant­ly.

Researchers and sci­en­tists have focused on unveil­ing secret mind func­tions since long. On the same foun­da­tion, I am pre­sent­ing the con­cept of sim­u­lat­ing the sub­con­scious mind with the help of Arti­fi­cial Intelligence.

Simulation of the subconscious mind

Sim­u­la­tion of the sub­con­scious mind would have all the same algo­rith­mic strengths and weak­ness­es as your own sub­con­scious mind. And, a very strong AI sys­tem will be able to see things that you aren’t able to see about your­self.

As I have afore­men­tioned, the unconscious/subconscious minds are unimag­in­ably com­plex, but what we are say­ing here is that there is a pos­si­bil­i­ty of build­ing an AI algo­rithm that could sim­u­late the process of a sub­con­scious mind.

But, before we jump into the point, we need to con­sid­er few more impor­tant points about the com­plex­i­ty of our brain.

DALL E 2 is cur­rent­ly hot and shak­ing the tech world with its abil­i­ty to con­vert any thoughts into image. This is a glimpse of the future we are going to enter. While DALL E 2 takes your text and con­verts it into art/design, DALL E 10 would be able to actu­al­ly design your sub­con­scious brain. By the time we reach DALL E 20 or some­thing equiv­a­lent to that, AI will be able to sim­u­late the sub­con­scious brain.

It’s really difficult to simulate our subconscious mind

Our sub­con­scious mind is our most pow­er­ful tool. It is the part of our mind that makes deci­sions with­out our think­ing about it at all.

Our sub­con­scious mind sees, hears and feels things we are not able to see, hear or feel due to lim­it­ed amount of atten­tion span. It decides how we feel and what we do with­out us con­scious­ly think­ing about it.

We know that decod­ing human behav­ior is an extreme­ly dif­fi­cult prob­lem in the field of AI research. As any data sci­en­tist can tell you, the process of under­stand­ing behav­ior and pre­dict­ing future actions is a very chal­leng­ing one, requir­ing high pre­ci­sion in order to pro­duce accu­rate results.

The nature of our sub­con­scious mind has been inves­ti­gat­ed by sci­en­tists for hun­dreds of years. The goal is to help us under­stand and con­trol our own un/subconscious behav­iors. But, it seems that our sub­con­scious mind is not so eas­i­ly under­stood as the HISTORIES of psy­cho­analy­sis — Fraud, Jung, Aserin­sky, Goodale and Mas­si­mi­ni — could not get any­where close to the total­i­ty of the most com­plex part of the human existence.

More­over, con­scious­ness in humans is often mis­un­der­stood as being some­thing sep­a­rate from the brain process­es we expe­ri­ence to pro­duce thoughts, feel­ings, and behaviors.

But, the con­scious­ness is a part of the brain itself — the brain’s phys­i­cal struc­ture and neur­al activ­i­ty may affect con­scious­ness. But they do not amount to a kind of human con­scious­ness that exists inde­pen­dent­ly from the phys­i­cal brain.

The Process of Human Learning and Memory

The process of human learn­ing is divid­ed into two main cat­e­gories: con­sol­i­da­tion and stor­age. Con­sol­i­da­tion refers to the sta­bi­liz­ing of new­ly-formed con­nec­tions or mem­o­ries while they are being stored in the brain. Stor­age refers to how new infor­ma­tion is trans­lat­ed from short-term to long-term mem­o­ry .

The brain stores mem­o­ries through a del­i­cate bal­ance of orga­ni­za­tion and flex­i­bil­i­ty. It can store new mem­o­ries effi­cient­ly because the con­nec­tions between neu­rons are strength­ened with use (an idea known as synap­tic plas­tic­i­ty). But at the same time it can also eas­i­ly reor­ga­nize its con­nec­tions when nec­es­sary (known as neu­ronal pruning).

As a neur­al net­work, our brain must obey the same rules for stor­ing and con­sol­i­dat­ing infor­ma­tion in both our short-term and long-term memory.

Natural language processing (NLP)

Nat­ur­al lan­guage pro­cess­ing (NLP) uses machine learn­ing mod­els based on neur­al net­works to esti­mate prob­a­bil­i­ties of word sequences giv­en a set of pre­vi­ous words.

The more neur­al con­nec­tions between neu­rons and the more train­ing data used in machine learn­ing algo­rithms, the bet­ter accu­ra­cy will the algorithm’s pre­dic­tions become.

A neur­al net­work can be applied to oth­er fields as well, such as rec­og­niz­ing faces and scenes in videos by ana­lyz­ing them using a series of fil­ters and regres­sion algo­rithms. NLP sys­tems can be used to extract fea­tures from voice record­ings and facial fea­tures to iden­ti­fy peo­ple or things in images.

Using a series of fil­ters, NLP sys­tems can sep­a­rate out acoustic fea­tures (such as pitch or vow­el for­mants) from speech into vec­tors or vec­tors that rep­re­sent each of the audio fea­tures. They can then use these vec­tors for fur­ther clas­si­fi­ca­tion purposes.

AI Simulating the Mystery of Subconscious Mind?

Yes. Arti­fi­cial Intel­li­gence of the next lev­el will prob­a­bly be able to decode the mys­tery of net­work of the sub­con­scious mind.

A neur­al net­work is a net­work of neu­rons that rep­re­sents a sim­pli­fied mod­el of the brain.  It is made up of var­i­ous lay­ers: input, hid­den and out­put lay­ers. NLP sys­tems use these neur­al net­works to process audio and visu­al inputs in order to pro­duce an accu­rate response.

One way in which NLP sys­tems can be used for this pur­pose is by exam­in­ing the rela­tion­ships between words.  A word’s mean­ing (seman­tics) pro­vides the most use­ful fea­tures for under­stand­ing its rela­tion with oth­er words (syn­tax).

For instance, if you want­ed to build an algo­rithm that could pro­duce metaphors, a neur­al net­work can use seman­tic and syn­tac­tic fea­tures to clas­si­fy whether a giv­en word or phrase was metaphor­i­cal or literal.

Tak­ing it to the next lev­el, AI can train a neur­al net­work to mim­ic the human mind. This works well to help under­stand what self-aware­ness is, and why we are so good at mak­ing decisions.

Here is a hypo­thet­i­cal ques­tion: If AI could build a sim­u­lat­ed sub­con­scious mind, what would it do? We’re not talk­ing about a sim­ple sim­u­la­tion here; we’re talk­ing about some­thing arti­fi­cial that could work in all of our every­day life.

A sen­tient Arti­fi­cial Intelligence(AI) — or Arti­fi­cial Arti­fi­cial Intelligence(AAI) — could poten­tial­ly repli­cate the human brain process by which behav­iors learn from expe­ri­ence as well as cre­ate new behav­iors, which is known as gen­er­al­iza­tion and abstrac­tion (the idea and def­i­n­i­tion of which will be dis­cussed more in the future). 

We can use neur­al net­works to sim­u­late the infor­ma­tion pro­cess­ing and deci­sion mak­ing process­es that are part of us. We may even be able to apply this mod­el to under­stand the caus­es and expla­na­tions for consciousness.

Brain, Subconscious brain, and simulation

In such, the brain itself, being a neur­al net­work(at least for this arti­cle), has the abil­i­ty to learn and adapt to stim­uli. If a com­put­er pro­gram is able to mim­ic the essence of our own neur­al net­work, which every neu­ron (or arti­fi­cial neu­ron) con­nects with oth­er neu­rons through synaps­es that allow for vary­ing strengths of sig­nals to be sent between them, then it would be rea­son­able to assume that such an algo­rithm could pro­duce sim­i­lar learn­ing results as our own minds.

Now, this sounds real­ly sim­ple on paper. But being able to sim­u­late a sim­u­lat­ed brain does not nec­es­sar­i­ly mean we can do it, at least not direct­ly. How­ev­er, if a neur­al net­work can learn and adapt to stim­uli, then it would be pos­si­ble to engi­neer one so that it can do the same thing in response to stimuli.

The brain is beau­ti­ful­ly struc­tured in an intri­cate way that allows it to be both com­plex and sim­ple at the same time. In order to sim­u­late an entire brain, we would need to sim­u­late every part of it. It must be sim­ple and com­plex at the same time too. The sim­plest way to get around this prob­lem would be to use more than one arti­fi­cial brain.

It might be actu­al­ly pos­si­ble to sim­u­late the sub­con­scious brain if we start with an arti­fi­cial neur­al net­work and add this addi­tion­al lay­er of abstrac­tion. If the sub­con­scious and con­scious minds are able to learn and adapt in response to stim­uli, then it would only make sense that we could do the same thing in response to stimuli.

In order to sim­u­late the sub­con­scious brain, which con­tains many dimen­sions, lev­els and lay­ers with­in itself that are not vis­i­ble, we would need an algo­rithm capa­ble of per­ceiv­ing extra-per­ceiv­able sens­es. Why? Because the sub­con­scious mind resides in a dif­fer­ent plane of exis­tence, from the point of view of our con­scious minds and our 5 senses.

So, if a neural network can help us better understand ourselves, what is the significance of this?

“The human mind, by its unique com­po­si­tion, is a com­plex sys­tem made up of bil­lions of inter­con­nect­ed neu­rons. Not only is the human brain not per­fectible; it seems to be impos­si­ble to repli­cate the var­i­ous ways in which it learns and changes as a result of expe­ri­ence and stimuli.

Each one of us is unique, and that unique­ness is essen­tial to under­stand­ing our­selves and the world around us.

A neur­al net­work could poten­tial­ly be an addi­tion­al and use­ful tool in assess­ing our minds and behav­ior bet­ter; how­ev­er, it will nev­er be able to replace our human mind. Our indi­vid­u­al­i­ty, mem­o­ries, expe­ri­ences, and feel­ings are what make us what we are. But, will we have to lose this all after AI starts decod­ing our sub­con­scious? I think we should dis­cuss on this ques­tion only after we are sure that AI has been that much capable.

Now, let’s see what AI needs to simulate our subconscious mind:

First of all, the AI needs to learn. To learn, it needs infor­ma­tion. So, it uses the same method that it will use lat­er to process data: It uses a machine learn­ing algorithm(MLA) to select which infor­ma­tion is more rel­e­vant than oth­ers; giv­en a pri­ori data, it can build its own mod­el (like a neur­al net­work). After this selec­tion and mod­el­ing of “rel­e­vant” infor­ma­tion, the AI uses rein­force­ment learn­ing to deter­mine what behav­ior should be select­ed next (for exam­ple: If you have select­ed a route and end up in an acci­dent; then you would prob­a­bly avoid tak­ing that same route again).

To per­form “future/desired” actions, the AI needs to sim­u­late the brain’s con­nec­tions and neu­rons. It uses a par­tial­ly con­nect­ed neur­al net­work with mul­ti­ple lay­ers of neu­rons. Since a par­tial­ly con­nect­ed neur­al net­work is not capa­ble of learn­ing new infor­ma­tion, it needs to be able to inter­po­late between two or more networks.

To sim­u­late the sub­con­scious brain, the AI uses a neur­al net­work (or a series of neur­al net­works) that includes sev­er­al lay­ers of neu­rons. It is called a deep learn­ing neur­al network.

In the final step, the AI will gen­er­ate an algo­rithm that sim­u­lates how we learn from expe­ri­ence and how we make deci­sions based on past events.

How will the simulation look like?

Well, it is a long, hard process but we can sim­u­late the brain as an arti­fi­cial neur­al net­work. The cre­at­ed algo­rithm then could become a sim­u­lat­ed ver­sion of our sub­con­scious men­tal algorithm.

It will look like a sim­pli­fied mod­el of our sub­con­scious work­ing to pro­duce the desired results.

We’ve had a glimpse into the future of AI: It will be able to sim­u­late what a human mind is and act as if it were con­scious. And at some point in the future, arti­fi­cial beings could even become con­scious – our great­est fear of what AI can bring to life, a real AI upris­ing.        

Real AI?

A fast-grow­ing top­ic in AI is that of the dif­fer­ences between AI and real, phys­i­cal AI. Real AI is made of sev­er­al parts, includ­ing hard­ware (which affects the num­ber of oper­a­tions an algo­rithm can per­form) and soft­ware (which con­tains all the rules that gov­ern its behavior).

There are some peo­ple who believe that human thoughts could be repli­cat­ed using AI. Oth­ers believe that there will always be a dif­fer­ence between nat­ur­al human intel­li­gence and machine intel­li­gence.  Final­ly, there are those who do not agree on this issue at all.

Why? If you read through our arti­cle thor­ough­ly, you’ll see how machines are slow­ly get­ting clos­er to achiev­ing human-lev­el intel­li­gence by mim­ic­k­ing the human brain.

In fact, AI has already achieved many of its goals, such as the abil­i­ty to learn how humans think and how we solve problems.

This is by no means a threat to our exis­tence – it’s just anoth­er tech­no­log­i­cal inno­va­tion that will help us under­stand real­i­ty and improve our lives. But it’s also impor­tant to be real­is­tic about these new tech­nolo­gies; there is still much work to do before arti­fi­cial intel­li­gence can per­form all of the tasks that a human brain can.

Finally,

Sim­u­lat­ing sub­con­scious brain would mean unlock­ing sixth sense.

We may nev­er be able to recre­ate a full human brain, but that doesn’t mean that we can­not build a syn­thet­ic con­scious mind.

A con­scious mind can be sim­u­lat­ed by cre­at­ing an arti­fi­cial neur­al net­work capa­ble of mim­ic­k­ing human infor­ma­tion pro­cess­ing and deci­sion mak­ing process­es. It could also include mem­o­ry and past expe­ri­ences stored in a database.

As you can see, there are a lot of things hap­pen­ing in the field of AI devel­op­ment, and much more, includ­ing an AI care­ful­ly decod­ing the sub­con­scious mind, are like­ly to hap­pen in the future. AI is slow­ly but sure­ly start­ing to gain con­scious­ness and intelligence.

It seems that the future of arti­fi­cial intel­li­gence is not only going to be inter­est­ing, but also adven­tur­ous in many ways. We cur­rent­ly get sat­is­fied with new things to feel, in the future it will be about get­ting new sens­es to perceive.

Leave a Reply

Your email address will not be published.

Join our NewsletterDaily Glimple of Future

Our blog, "Daily Glimpse of Future", strives to make the future much clearer than it is today. Join our newsletter for free now.