By Zita Perera Subasinghe
No, it’s not Zita preaching doom and gloom, but this conundrum has been bothering humans in the last
few decades and books have been written about it. Films have been made, and
they’ve been phenomenal box office successes. Remember AI?
The Industrial Revolution in the past nearly two
centuries made man resort to the ease and speed of machines to do the hard and
heavy jobs. Fine! Why shouldn’t they?
But with the birth of information technology and its
rapid expansion in the latter half of the last century onwards, and its
phenomenal success in the last two decades, made it possible for a new
generation of machines and especially computer technology to surpass man in
efficiency, speed and the intricacy and complexity of what they are capable of
doing.
Machines for leisure, fitness, and machines for
cooking and automobiles for speedy, efficient transport have been with us for
at least a century. Now we are looking at self-drive cars!
Aren’t we worried about climate change? Well, a new
breed of machines focusses on rescuing the planet’s ecosystem, and they’ll look
upon concerned humans with benign disinterest.
Sounds like science fiction? No! And there’s going to
be autonomous weapons systems which do not need human intervention. And they
are not just going to be ‘terminators’. They will be able to discriminate right
from wrong, friend from foe, and make a good judgement on whether it is assailant
or opponent justly and humanely and take action just like a human soldier
would.
Our relationship with technology is changing! We can
conduct a basic conversation with Alexa about day to day matters, and Siri will
answer our IT questions and give advice.
One problem though! We have potential surveillance devices in our homes.
Drones are to be used to detect speeding motorists
even if our GPS helpfully warns us ‘Speed check drone ahead’! Invisible eyes
are watching us! Are our digital servants turning into digital masters?
Well, it is up to us to some extent. And they will
never take over humans in compassion, love, companionship, forgiveness and oh
yes, creativity! A human being easily surpasses a robot or drone in these
qualities,I like to think!
Wow! This will make us all pause and think, think and pause. I doubt whether anybody can accurately predict the future in general and the relationship between Man and Machine more specifically. What we know is that the amazing abilities of Man has continually sought to discover ways and means of satisfying its desire for prolonging life, making it more pleasurable and using aids such as Man-made machines to help achieve this. Whether it is a car, or an aeroplane or mobile phone or MR scan, the drive is the same. Man is trying very hard to "delegate" as much as possible and apart from the emotions that I described, others such as the desire to be more and more better off materially makes them seek profit and money, and this has been a big driver, especially in Capitalist Societies. The worry now is that we are able to make Machines solve problems (in certain domains only at the moment- for example, AplhaGo Zero has totally surpassed the best human experts on the game Go ), far more accurately and far more quickly than we can. There is therefore a growing fear that as we invent Machines which are now capable of "learning", these will have more powers which ultimately may even learn baser human emotions such as exploitation, hatred, animosity, anger and ultimately direct it at ourselves, their main competitor!
ReplyDeleteI do hope a balance will be reached and Man and Machine will complement each other harmoniously.
It is salutary to reflect on the fact that we share 96% of our genes with Apes but that extra 4% makes all the difference! Will machines be ever capable of matching the entire range of human capability? We don't know. What we know is that the analytical capability of computers is growing exponentially and in terms of crude "intelligence", they will surpass us, but,we have something more, that indefinable "something", may be "consciousness","soul","mind", that makes us swoon at a beautiful painting or a mood elevating Beethoven Sonata!
Mahendra, thank you for the above comment, of which a couple of points are very reassuring, in this debate. That is: machines can be smarter and your example of gaming machines like in the game Go, and of course as we also know, in chess, it's fine as far as it is for entertainment. But if they learn baser emotions this won't be a happy place to live in. There again, I think it's in man's hands to stop making these machines before it reaches that stage. And I echo your sentiment that Man and Machine will complement each other! I am glad you made the example of Beethoven Sonata. We cannot ever envisage a day when animals can overtake Mozart or Beethoven! Creativity, music, higher thinking, meditation are thankfully the privilege of human beings. Zita
DeleteAt present the benefits of technology far outweigh the drawbacks. We have the ability to keep this under control. For some it is Artificial Intelligence that cause them concern like creating humanoid robots. It is important to evaluate the benefits and the dangers to avoid a disaster.
ReplyDelete“Everything we love about civilization is a product of intelligence, so amplifying our human intelligence with artificial intelligence has the potential of helping civilization flourish like never before – as long as we manage to keep the technology beneficial.“
Max Tegmark, President of the Future of Life Institute
We should worry more about intolerant political ideologies, Climate change, antibiotic resistant pathogens and nuclear weapons before artificial intelligence and the forward march of technology that at times seem intrusive.
“This is a mournful discovery.
1)Those who agree with you are insane (paranoid)
2)Those who do not agree with you are in power.”
― Philip K. Dick, VALIS
I don't know whether our friends have come across Sadhguru. He is an Indian Yogi and is full of wise expositions. He welcomes AI and says it will be wonderful to be freed from so many mundane tasks just as machines in other areas have done. Here is a link if you fancy watching him.
ReplyDeletehttps://www.youtube.com/watch?v=w0wJbU6QEgQ
It is so interesting to hear the arguments of Nihal in giving artificial intelligence the right place and importance and not elevate it above human capabilities and in this respect I do agree that the negative and intolerant ideologies of humans give us more reason to worry about. We do find that wild animals stay in the wilds and if we leave them alone they rarely go out of their way to attack us. But not so with humans. We only have to think back on 9/11 and the last Easter Sunday attacks. I agree, we should only try to control human aggression rather than that of artificial intelligence. Zita
ReplyDeleteZita, I like the optimism with which you write.
ReplyDeleteThe picture that accompanies the post is also one to be taken note of - a situation we need to avoid. The day when machines will be able to “discriminate right from wrong, friend from foe, and make a good judgement on whether it is assailant or opponent justly and humanely and take action just like a human soldier would” is a little way off. We might see this day if we are lucky to survive what the machines are programmed to do by man.
The concern now is the developement of lethal autonomous weapons (LAWs) which are programmed to kill by algorithm - no feelings or right or wrong here !
I shall try to insert here if I can - two articles which every thinking human being has to read, regarding where we are at with AI right now and it’s serious implications .
We do not need to fear about machines taking man over - The evil that man is about to resort to with machines as they are might see us off the planet before we see that day. The articles are- “The Truth About Killer Robots - Wikipedia”, and
“Death by algorithm: the age of killer robots is closer than you think”.
Sorry to be a wet blanket - let’s be consoled by the last para in your post.
Cheers
Technology brings you great gifts with one hand, and it stabs you in the back with the other. C.P. Snow
ReplyDeleteYou always bring to our notice some extremely valid points and you give us references. This one you've given, i.e. The truth about killer robots' somehow is impossible to find, at least I wasn't able to on Wikipedia. But even without reading it I accept that the day when machines will be able to 'discriminate right from wrong' is a way off. Just like all other inventions must have gone through this problem until the inventors and users learnt to use them ethically and laws have been produced and enacted to prevent inventions being used for unlawful purposes, AI will also have a test period where there will be controls in place, at least I hope that's how things happen. This brings me back to a previous argument that we often have to give in to a lesser evil to prevent a greater evil. Here, the 'evil' merely becomes non acceptance of an invention.' If by that more people don't get the help they need then we have to use the principle I am talking of. Do I make sense?
ReplyDeleteZita, Of course you make sense !
ReplyDeleteAs has happened in the past with treaties aimed at preventing nuclear proliferation,testing , and promotion of nuclear disarmament etc, as well as measures by the CBDE to outlaw chemical warfare, the UN’s CCW has been trying to ban manufacture of these lethal autonomous weapons- without success so far with a handful of powerful nations resisting the move. One ‘power’ even said they will not use them, but will not stop manufacturing
them ! Ethical do you think ? !
I get what you say about choosing the ‘lesser of two evils’
I am all for advancement of AI for its use in helping humanity as already widely seen in delivery of medical care specially to remote areas , in search and rescue operations etc but Iam against it’s use for developing autonomous weapons.
It is not the advancement of AI itself, but it’s use for evil purposes by ‘man’ which is a concern.
Iam sorry you couldn’t access one of the articles - I shall try and send it to you. Cheers
Here is something more to worry about. There is a new technology called Multi Agent Artificial Intelligence (MIAA) which is capable of making very accurate predictions on scenarios using simulated populations. The populations use what are called “agents” which are virtual members of the population being simulated but take into account many more factors rather than the usual age/sex, affiliations etc but use a range of behaviour patterned agents in complex patterns as you would find in real life. This is now possible because the enormous computer power required is now available. It can be used to predict political outcomes as well and it is said that Trump agents would almost certainly use this technology as it is capable of giving predictions according to various induced behaviour patterns in the “agents”. This would enable them to choose the most effective political message and also target propaganda even better than they do with current models.
ReplyDeleteIt was used for the benefit of Man in the Ebola epidemic where modelling predicted that left unchecked, it would affect 1.4 million people. It was then used to test various interventions and the decisions made were arguably responsible for the fact that infections being restricted to 28.000 people.
Here is the worrying part. It was used, just as a modelling exercise, to see what is the most powerful intervention that could be made on a hypothetical population with a majority religious group In conflict with a minority one (because such societies easily spiral into deadly violence). The simulation was run and came up with an answer for the most efficient way to restore peace – it was genocide! This shows the importance of ensuring that the guidelines for modelling or the rules that AI develops to provide a solution are managed carefu;;y and ethically. This will be increasingly important as such models will be used more and more for scenario planning. In the example given, purely on a dispassionate “machine”-like “thinking”, genocide was the most efficient option. I can see misuse of such technology being used in Medicine in choosing the most cost-effective medical solution for disease related conditions where the “agents” are de-humanised. Hopefully, that won’t happen.
This exactly is what concerns me too.
ReplyDeleteWith the ever increasing capabilities of AI, ‘man’ will programme the machines whichever
way they can to achieve their evil outcomes. Ethnic cleansing and Genocide are real possibilities in areas of conflict- hopefully not in medicine.. There is much on this online, including comments by some ‘world leaders’ - too political to write here !
Indeed! A very sensitive subject that needs discussion, but our blog is probably not the place to go into details. But awareness of the benefits and dangers of AI need to be more widely known. I have sufficient confidence in Humanity that the right balance will be achieved. The awesome power of humans to shape the environment they live in is and has always been intriguing.
DeleteMahendra, I read with interest what you say. I have to admit that 'Multi Agent Artificial Intelligence' is something I have not yet heard or read about. And this is typical of the misuse of AI by humans to achieve their own evil ends. AI does not have a conscience or emotions so if man puts AI into use to justify man's own evil ends then that should definitely be forbidden. I expect more and more regulation and control by governments will be and should be introduced as this field advances and man uses it as a means to rationalise his own abominable behaviour. This again emphasises the premise that 'man can be more evil than machines'. And it does not take anything away from the wonderful, useful, quicker and beneficial uses of Artificial Intelligence. Zita
DeleteZita, I can understand your concerns but so far the scenario I mentioned about ethnic balance has not been used,at least not overtly. But the great thing about AI is you can introduce rules which it then obeys. If the scenario I mentioned is run with "genocide" not an option, other measures will be suggested. AI is still not completely autonomous as far as I know. The Ebola example was one of good use by the way. I think we just have to be careful. So far no terrorist group has got hold of a Nuclear Bomb which has the capacity to annihilate mankind. I feel that AI will be regulated but things that happen under the surface are more worrying than the extreme example I cited I am referring to all of us being manipulated by the Media and the Media Moguls will not hesitate to use sophisticated AI technology to influence our voting habits, our buying habits, and all sorts of other behavioural features to sell us what we don't need and to vote for people we don't want!
DeleteMahendra, I have to say in great admiration, that you have brought out some truly relevant and little known facts in this man vs machine (i.e. technology) debate. It is very important to us to be aware of these dangers which seem to be already here and you have talked about 'things that are under the surface'. I have been listening also that machines will replace man in all news broadcasts all over the world! And News will be also be 'gathered' by robots! Can you imagine that? What if robots decide to spread 'fake news'!
ReplyDeleteWell, these things are already here. There is no escape. But one consolation is that man can be more evil than any machine. But man can also be kind, gentle, loving and forgiving. These things are unknown to machines but can be taught as time goes on maybe but then there is no spontaneity. Well, we can go on! Thank you again, Mahendra, for all the fine points you brought forward in this discussion. Rohini's and Nihal's contributions on this subject too were, as usual, very interesting and enlightening! Thank you all!
Zita, Mahen, Nihal and others who are interested-
ReplyDeleteOn the other side of the coin of AI , Elon Musk’s progress creating a symbiosis of machine and man at Neuralink is a worthwhile read and holds promise in managing numerous neurological conditions, but also makes one wary of their use in the distant future to re-programme human minds ! This can be seen as just an extension of the current use of cochlear implants and Deep Brain Stimulation In Parkinsonism etc
The full story which is rather long boggles one’s mind ! However an easy-to-read link follows-
https://www.dezeen.com/2019/07/22/elon-musk-neuralink-implant-ai-technology/
shall check it out Rohini. Thanks. Mahen
Delete