Can AI system be modeled on the human brain?

The mind is an incred­i­bly com­plex enti­ty, so much so that even the most pow­er­ful arti­fi­cial intelligence(AI) in existence—the human brain—isn’t ful­ly under­stood. It remains a feat of immense intel­lec­tu­al and tech­ni­cal prowess, and is still the most pow­er­ful intel­li­gence ever cre­at­ed. So, what would hap­pen if we tried to use our under­stand­ing of the human brain (or even some aspects of it) in order to get mod­eled an AI sys­tem that could approx­i­mate how a nor­mal per­son behaves?

That’s the ques­tion this post is going to explore. Hope­ful­ly, it will help you make an informed deci­sion about whether try­ing to emu­late a human mind for your own AI sys­tem is some­thing you want to pur­sue or avoid.

Firstly, what does getting AI system modeled on the human brain mean?

AI system modeled on human brain

It’s impor­tant to under­stand that get­ting an AI sys­tem mod­eled on the human brain does­n’t mean that you’re try­ing to turn your AI sys­tem into a human (although it might make it appear more human-like).

You should keep in mind that the brain isn’t even real­ly a com­put­er. It’s but rather a series of incred­i­bly com­plex cir­cuits and algo­rithms that respond to inputs. It’s far beyond any con­ceiv­able tech­nol­o­gy we could cre­ate, so attempt­ing to mod­el the mind of that very advanced process is just silly.

Rather, mod­el­ling an AI sys­tem on the brain means that you’re try­ing to use the brain’s cir­cuits, algo­rithms, and struc­ture as a guide or tem­plate. You’re try­ing to emu­late some of the process­es involved in the brain’s oper­a­tion, so that you can emu­late some of the process­es involved in human thought.

That’s why a lot of peo­ple think that using an AI to mod­el how we think is a good idea — both from an intel­lec­tu­al curios­i­ty stand­point and from an AI devel­op­ment stand­point. They don’t want to cre­ate any form of machine intel­li­gence that isn’t human-like. They want their sys­tem to exhib­it the same intel­li­gence that humans exhib­it when they’re thinking.

Douglas Hofstadter’s idea about getting the human brain modeled on an AI

The most famous pro­po­nent of imple­ment­ing an AI with a human-like mod­el is cer­tain­ly Dou­glas Hof­s­tadter. His book, I Am a Strange Loop, is one of the most well-known books in this field.

His idea isn’t nec­es­sar­i­ly to copy how our brain works, but rather to try and under­stand and emu­late the sys­tem that gives our thoughts their form and struc­ture. He even argues that try­ing to mod­el the human brain on an AI is like­ly more ben­e­fi­cial than try­ing to design an AI sys­tem using our under­stand­ing of the brain as a guide.

At the very least, he argues that try­ing to under­stand our own think­ing process­es is a lot more inter­est­ing than sim­ply mak­ing an AI sys­tem that behaves similarly.

Our thoughts are much more com­plex than any AI sys­tem could pos­si­bly be able to sim­u­late. But that does­n’t mean human brain can’t be mod­eled on an AI system.

The advan­tage of get­ting an AI sys­tem mod­eled on the brain is that you can eas­i­ly iso­late, char­ac­ter­ize, and repro­duce these process­es. So a sys­tem cre­at­ed by repli­cat­ing these struc­ture and func­tions of the human brain would be a lot eas­i­er to debug and ver­i­fy than an AI sys­tem in which it’s impos­si­ble for humans to repli­cate how we think.

In fact, under­stand­ing how our brain works may even help us dis­cov­er new ways to improve it. We might be able to use our under­stand­ing of how we think in cre­at­ing bet­ter AIs.

What problems might you run into if you attempted to get AI system modeled on the human brain?

AI system on human brain
  1. The first major prob­lem you run into is an eth­i­cal con­cern. The human brain isn’t some­thing that can be eas­i­ly repli­cat­ed using today’s tech­nol­o­gy; there’s no way we could recre­ate the brain in a way that would­n’t have adverse effects. Even if we had the tech­no­log­i­cal capa­bil­i­ties, it would be uneth­i­cal and immoral to repli­cate it in any fashion.

As I men­tioned, the human brain isn’t easy to under­stand and repli­cate (it’s one of the most com­plex things in exis­tence), but we might try to get around that obsta­cle by sim­ply try­ing to clone a human brain. It seems unlike­ly that any­one would be able to do so, but it’s possible.

  1. The oth­er prob­lem is that it’s unlike­ly you’ll know how to get a human brain cor­rect­ly mod­eled in any man­ner that won’t result in fail­ures. Part of what makes the brain so com­plex is that it’s con­stant­ly chang­ing and adapt­ing. Because of this, it would be incred­i­bly dif­fi­cult to cre­ate a sys­tem that accu­rate­ly reflects how the brain works — at least, it would be dif­fi­cult com­pared to how easy the brain is for humans to under­stand and replicate.
  2. Last­ly, it’s also extreme­ly expen­sive to mod­el the human brain cor­rect­ly. That’s because it’s basi­cal­ly impos­si­ble for humans to repli­cate how the brain func­tions. It’s much cheap­er to use sim­u­lat­ed intel­li­gence than human intel­li­gence, since we know how to sim­u­late an AI sys­tem prop­er­ly and then let it evolve and devel­op on its own.

But still: Can AI system be modeled on the human brain?

Well, the one word answer for this ques­tion: Yes!

In fact, there are a few rea­sons to believe that we can cre­ate AI sys­tems that mod­el how we think. It’s not nec­es­sar­i­ly a very easy process, but it’s possible.

Most of our thoughts don’t involve a lot of plan­ning. We’re not try­ing to form com­plex plans for the future; most of our thoughts involve com­mon sense and sim­ple rea­son­ing — things that robots could mod­el fair­ly easily.

Our thoughts are much more com­plex than any AI sys­tem could pos­si­bly be able to sim­u­late. But that does­n’t mean human brain can’t be mod­eled on an AI sys­tem. We can come up with a rough approx­i­ma­tion of how we think, which is great because it’ll be rel­a­tive­ly easy to debug and verify.

Relat­ed Post:

We might not be able to cre­ate a per­fect copy of the human brain. But, that does­n’t mean an AI sys­tem can’t mod­el how it works in some form. That’s still enough to han­dle most of the jobs that AI sys­tems need to perform.

I’m not say­ing that the human brain is easy to repli­cate, but it’s cer­tain­ly pos­si­ble. And that might be enough for us to cre­ate a sys­tem that can learn from us in a sim­i­lar way to how we learn from each oth­er.

Clear­ly, it’s a dif­fer­ent way of get­ting the mind mod­eled; but it works for the pur­pose of learn­ing and devel­op­ing AI sys­tem. So it seems like a real­is­tic way to get the human mind modeled.

What’s next for human-like AI?

Right now, we’re still a long way away from cre­at­ing an AI sys­tem that can tru­ly under­stand the way humans think. But that does­n’t mean it won’t hap­pen in the future. It’s quite pos­si­ble that with­in our life­times, we’ll devel­op a sys­tem that can sim­u­late how humans think.

Rec­om­mend­ed: What’s the most human thing that a non-human intel­li­gence could do?

The rea­son I say this is because there’s so much research and devel­op­ment going on right now — and so many peo­ple try­ing to advance the tech­nol­o­gy — that get­ting AI sys­tem mod­eled on the brain of a human being seems likely.

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.