AI: A FRIENDLY PET OR A DREADFUL THREAT?


.                     (Image Source: Forbes.com)

Some sparrows decided they wanted a pet owl. Most of the sparrows thought the idea was awesome, but one was skeptical, voicing concern about how could the sparrows control an owl. This concern was dismissed in a "we'll deal with that problem when it's a problem" matter.
That’s how acclaimed Swedish philosopher Nick Bostron analogizes the present relation of Humans with the idea of AI, in his book Superintelligence.

Now, whether the “Owl” would prove to be a useful pet or a dreadful threat to the “Sparrows”, is a matter of intense debate. Some of the greatest leaders and AI scientists, in themselves, seem to be holding contradicting opinions.


In contrast to this is Mark Zuckerberg's opinion that AI is going to make our lives only better in the future, and  the statements about doomsday scenarios are “pretty irresponsible,”. Similar is the opinion of Jeff Bezos who says- “The idea that there is going to be a general AI overlord that subjugates us or kills us all, I think, is not something to worry about. I think that is overhyped,”


And Bill Gates sees AI as a positive labor replacement tool, writing that an AI revolution “should be positive if we manage it well.”

Different leaders sharing different opinions- obviously confusing for us commoners.

But all such opinions about future relationship of Humanity with AI are not exactly in the same context...
.                       (Image Source: Forbes.com)

Basically, when Dr.Stephen Hawking warned us about AI taking over the entire human race, he was talking about ASI (Artificial Super Intelligence) and when Jeff Bezos said that it is nothing to worry about, he was speaking about AGI (Artificial General Intelligence).

 
 What’s the difference?

AI has three stages of development-

ANI (Artificial Narrow Intelligence) - Also known as 'Weak AI', In simple terms, ANI is capable of performing a single task which it is programmed for, exceedingly well. But cannot generalize that capability to perform any other tasks.

AGI (Artificial General intelligence)- Also known as 'Strong AI', AGI refers to human-level intelligence i.e ability to perform any task that human Brain can like reasoning and flexible thinking combined with computational abilities of a computer. 

ASI (Artificial Super Intelligence)This is a matter of concern for the world as ASI means an AI that is better than humans in every single task. Basically- the AI from Sci-Fi movies.


HOW FAR ARE WE FROM THE AI FROM SCI-FI (or ASI)?

On March 9, 2016, AI-enabled ‘AlphaGo’ defeated the world champion in the ancient game of Go. Taking the world by surprise. Go has more combinations of moves than the number of atoms in the observable universe! AI learned it and came up with its own moves that even the experts had never seen. It is regarded as a milestone moment in the development of AI.

In 2017, it became a matter of heated gossip and debate when Facebook had to shut down its ambitious AI experiment when two AI systems started communication between themselves in their own ‘secret’ language, only understood by them!
.                       (Image Source: Mirror.com)
While incidents such as above might create a sentiment of a soon approaching AI dominance among the people, experts say that we still have decades to go before we realize even the most basic version of Artificial General Intelligence. 

That's because the system that beat the best human players in the complex game of Go was REALLY good at… playing Go. It is not able to take that incredible expertise and apply it to ANY other problem. This is why the best learning machines don’t actually think. 
And in contrast to the media hype, the AI robots that Facebook had to shut down, could only communicate regarding what they were programmed for - negotiate on the price of provided sample goods like fruits and toys!

Both are examples of some of the best AI tech we have today. But their capabilities are limited to expertise in their own single domain.

For example, a human who can read Japanese characters would likely understand Japanese speech, know something about Japanese culture, and can even make good recommendations at Japanese restaurants. In contrast to this, very different AI systems would be needed for each of these tasks. 

If a machine can acquire knowledge in different fields and apply it across all, just as humans do to survive in the real world, only then it can be called Artificial GENERAL Intelligence.

 Human brain neurons are not ‘special purpose cells’. They can change and adapt according to different purposes like the same neuron that is activated while crying, gets activated when we are lying to someone. Unlike the Artificial Neural Networks used in AI systems that can only perform the single task, they are designed for.
                 (Image Source:kdnuggets.com)
                    (Image Source: favpng.com)

The human brain has over 100 billion neurons connected in many interesting ways to make over 100 trillion neural connections in the average human brain.

This combined with the fact that we are yet to fully decipher the functioning of our own brain, it indeed seems like we are nowhere close to achieving AGI.

However, the opinions are held by the world’s scholars regarding the time required for achieving AGI range from less than 11 years, till 2045, till 2099, to Never.


So, How ‘Intelligent’ Is today's AI?

 AI is presently revolutionizing health care and warfare; it is helping people make music and books; it is scrutinizing people’s resume, used for Fintech applications like judging people’s creditworthiness, making paintings and constantly working to simplify our tasks by being present in our phone. In short, it is influencing our lives in all possible ways.

But in real terms, even the most advanced A.I. today is, in a way, just an intelligence-mimicry. For instance, Google search churns out billions of queries in merely miliseconds, but it’s not fully encompassing sentient intelligence i.e. Google does not understand a human-like meaningful interpretation of the entered query, rather the results are just the outcome of merely some algorithmic formulations in the back-end.

IBM's famous Watson AI recently showed an excellent performance in the complex questionnaire game called Jeopardy.But it doesn’t mean that it really understood any of the questions it was asked in the way a human might. 

No current computer or any type of A.I. can be said to have the least amount of sentience (capacity to feel or perceive) compared to any living organism on the planet.




THE REALISTIC CONCERN WITH PRESENT AI-

While the concepts of Artificial General Intelligence and Artificial Super Intelligence are pretty unrealistic in the present, the current AI is capable enough to have raised a real concern for humans-- JOB LOSSES DUE TO AUTOMATION. 

Elon Musk recently raised a concern that soon governments might have to fix a universal basic income for all as automation might leave almost no jobs for humans. A recent McKinsey report states that close to 50% of all the current work activities are technically automatable using the present AI.
A recent Economic times report states that, at present, only 10 jobs are created for every 100 jobs taken away by Artificial Intelligence .

These stats raise the question of humans staying unemployed and AI taking over all our jobs. But that is only half of the story!
Though AI is destroying previous jobs, it will create more new jobs than it destroyed in the future, as this CNBC report says.


                            (Image Source: CNBC)
Industries know that while AI is better than humans at some tasks, but does not have all the abilities required for a job better than humans.

For example, for building a bridge over a river, an AI system can come up with many possible designs based upon the constraints given to it. However, not every design it proposes would be practically feasible. Here a human civil engineer would decide the best possible solution according to various real-world factors. But surely, what would have taken a lot of time and effort can happen in much less time.

Hence, it is safe to say that most jobs would be affected by AI resulting in a collaborative working environment between us and AI.

WHAT CAN AI NEVER EVER DO?

CONSCIENCE-MORALITY-CREATIVITY: Not AI'S cup of tea--

Today AI can create paintings like Picasso by analyzing his style of painting through scanning his previous works.
So we have got our next Picaso!!! 
Have we?

                (Image Source: Bloomberg.com)
A PAINTING MADE BY AI

.                    (Image Source: Wikipedia)
A PAINTING MADE BY A HUMAN

Some people argue that now since AI can come up with new music compositions, new paintings, new designs, etc. It qualifies to be called as ‘Creative’. 

Let’s compare two paintings- the human Picasso painted vs the- AI Picasso- painted.

Human Picasso: Painted when he was so INSPIRED by some EXPERIENCE in LIFE that he couldn’t resist to paint his HEART out to COMMUNICATE his FEELINGS to us, fellow humans.

AI Picaso: Took numerous paintings of Picasso as data, analyzed them, and computed the variables and constants between them and derived a pattern using its complex algorithms. And painted (or printed?) a new painting that resembled the style of Picasso.

Which one would you like to hang on your wall?

The term ‘creativity’ has many definitions depending upon whom you ask, but if it has got something to do with human feelings, we humans can certainly not accept something as a piece of art if we know that it is fundamentally just a result of some complex mathematical computations rather than someone’s genuine feelings.
Wait, can a machine ever FEEL?
Well, they can make it ‘FEEL’ using data and algorithms.
Genuine Feelings?
Philosophers propose the famous “Chinese room” thought experiment to prove this point.

While opinions vary over whether machines would ever be able to feel, another question that arises is of MORALITY.

Our Morality is our perception of what is "Good" vs what is "Bad". We, humans, are yet to settle upon our own definitions of morality. What is Moral for you can be Immoral for me and vice-versa.
Provided this, how do we define morality for machines.

.                 (Image Source: aberdeen.com)
If a self-driving car finds itself in a situation where a collision is unavoidable, should the car knock down an old person in his sixties or a child? What if the child had a terminal illness? What if the old person was the sole caretaker of his partner? Who to hold accountable if someone dies?

Even though some scholars argue that even AI will have its own conscience and would be able to think creatively and ethically in the future, It does not seem quite possible in the foreseeable future, considering the fact that we are yet to completely understand the cognitive functioning of our own brains!

It is due to these very reasons that some jobs actually show an increase in employment despite AI automation. They are- Human Resource Managers, Sales Managers, Marketing Managers, Public Relations Managers, Chief Executives, Event Planners, Writers, Software Developers, Editors, and Graphic Designers, etc.

If observed closely we can see that all these jobs demand creativity or a high level of Emotional Intelligence. So, it is a clue that if you apply creativity with Emotional Intelligence in whatever job you do, you are doing something impossible for AI. And hence AI can't replace YOU.

A LESSON FROM HISTORY-
September 26, 1983. The computers of the Soviet Union's defense system started flashing warning signs of a nuclear attack initiated by the West. A mere 10 minutes to act, while the screens showed 5 nuclear missiles heading towards the country.

Colonel Stanislav Petrov was confronted with a decision that could shape the future of humanity. He could 'retaliate' according to the military 'training' given to him or, trust his cognitive rationalities that if the US was to wage a war, why would they do it with just five missiles? He speculated that the system was new and not fully trustworthy yet and was aware of the responsibility on him that given the world's political scenario of that time, a single ill-informed mistake was enough to ignite the Third World War. 
Needless to tell, He ignored the warnings.
And it turned out that the warnings were indeed false- caused due to a flaw in the systems.

Had an AI been controlling the defense system, what do you think would have happened ?




CONCLUSION

It is pretty clear that the Sci-Fi state of AI taking over the human race is out of the question in the foreseeable future.
But the real concern of AI taking over most of our jobs is surely a matter of worry to some of us. But if we look at the jobs that are at the least risk from AI automation, all these jobs need creativity or high Emotional Intelligence.
 So, it is a clue that if you apply creativity with Emotional Intelligence in whatever job you do, you are doing something impossible for AI (AI seems to be giving the lost lessons of honesty to humans :-). And hence AI can't replace YOU. 
And if changes are the only constants in this beautiful world, optimism is the best perspective to welcome them with.

IF WE ARE ABLE TO REGULATE AI WELL,    WE CAN SEE A SOCIETY WHERE ALL NON-CREATIVE/ITERATIVE TASKS ARE PERFORMED BY MACHINES, GIVING HUMANS MORE TIME AND ASSISTANCE FOR THEIR CREATIVE  ENDEAVOURS. IT MIGHT HELP US ERADICATE POVERTY AND TREAT THE DEADLIEST OF THE DISEASES. MAKING THE WORLD A HAPPIER PLACE.

RESOURCES:

https://ai.stackexchange.com/questions/2111/is-ai-living-or-non-living
https://www.zdnet.com/article/what-is-artificial-general-intelligence/
https://www.bbc.com/news/technology-30290540
https://www.cnbc.com/2018/05/11/jeff-bezos-on-ai-robots-wont-take-all-our-jobs.html
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwjpoZr1q5_qAhX_8HMBHQaBAOcQFjAAegQIAxAB&url=https%3A%2F%2Fwww.cnet.com%2Fnews%2Ffujitsu-supercomputer-simulates-1-second-of-brain-activity%2F&usg=AOvVaw1TV0uGNPBPc9KSolSgJStX
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwjGhOmrrJ_qAhUowTgGHVXKDyEQFjAAegQIAhAB&url=https%3A%2F%2Femerj.com%2Fai-future-outlook%2Fwhen-will-we-reach-the-singularity-a-timeline-consensus-from-ai-researchers%2F&usg=AOvVaw1ynKpJ-YLldSGev_gW6NYx
https://www.theverge.com/2018/8/13/17670156/deepmind-ai-eye-disease-doctor-moorfields
https://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=&cad=rja&uact=8&ved=2ahUKEwj_0O7ArZ_qAhXJ7XMBHYfbB6sQFjAAegQIAxAB&url=https%3A%2F%2Fwww.nytimes.com%2F2011%2F02%2F17%2Fscience%2F17jeopardy-watson.html&usg=AOvVaw1PJVlRKFdON8Somf9ZDYar

No comments:

Post a Comment

Thanks for Commenting !!!
We Value and Appreciate YOUR Feedback.....
Cheers.....