eSignature Lawfulness for Business Ethics and Conduct Disclosure Statement in India

How to eSign a document: eSignature lawfulness for Business Ethics and Conduct Disclosure Statement in India

a very good evening to everyone uh thank you so much for joining this session i am akshay manker i am an associate in loss eco today we have brought to you a very interesting session uh which is based upon an area which is not very much developed today we'll be talking about uh the alternate intelligence uh technology and i'll be introducing the host uh very soon today we are joined by mr shikhar gupta who is a research fellow and a phd scholar in delhi university he has an experience of researching in the area of ai artificial intelligence i'm sorry i i missed mispronounced it earlier so we'll be talking about liability on for misconduct we are artificial intelligence i'll i'll hand over this session now to mr schicker and we will see how what are the nuances of this area how how liability can be imposed right so i won't take any much more of your time thank you so much all to join in uh shikhar sir please take this session ahead thank you and my thanks to the team at los eco for inviting me to conduct this webinar on this emerging issue of artificial intelligence i welcome all the participants present here both students as well as legal professionals who have taken their busy time out to attend this webinar which i'm sure will be an enriching experience for both you as well as me i say me because the issue the topic today of artificial intelligence particularly the liability issues there are no concrete legal or regulatory frameworks regarding the topic in india and neither in most nations of the world so this webinar could not be a one-sided lecture but an interactive participation where you will gain from my insights and my research on the topic while i will also gain from your valuable insights in my uh that may contribute towards my future work in my phd so i don't want this webinar to be a one-sided lecture because that way it would not succeed so i want you all to be ready to participate at any moment and be ready to question or present your comments whenever you like and i i assure you that this webinar is not a very technical webinar it is a i will make you understand things in a very uh layman approach so that each of you would be able to present your views or any ideas suggestions or even good questions that may that may come as we go on through the webinar so let's start with the webinar today all right okay okay let me just uh get in here for just for a minute so before starting i'd let you know that there is this q a tab here just beside participants you can see your you can put you up your questions there if you want to make any contribution you can put up your uh you know opinions on chat as well you can comment here and uh you know just just uh let's make this session interactive uh to start with i'd really uh like to know which all areas everyone is coming from if you're a lawsuit student just write law student in in the chat box if you're a lawyer right lawyer if you're coming from any other area just just post uh you know where you're coming from let's let's make this interactive i'm waiting for people to you know put the things in check just write lawyer or law student in the chat box law student yes we have law student law student law student any any more responses anyone who is not a law student a lawyer or coming from any other area right so uh today we are mostly joined by law students so that is our today's demographic i'm sorry i had to get in but uh please please take a head so akshay even the legal professionals and the lawyers are law students for all of their life even through if you become professors lawyers professionals policy influencers we will have to be students because this field this discipline demands that these students remain students through all of our lives so i welcome all the students present here i am also a lawyer student i consider myself also a law student and there i see a comment that there are also lawyers here which again i considered as law students so let's start with my presentation here so akshay i hope my presentation is visible yes so the topic for today is liability for misconduct via artificial intelligence in grief i will tell you the issues that we are going to discuss today [Music] we that means that we will see artificial intelligence from the point of technology as well as what it should mean in law because artificial intelligence is a very broad term so we will attempt to narrow down the scope of artificial intelligence and now then we will see what possible kinds of concerns or misconducts may arise by the use of bi systems and possible ways to tackle those thirdly we will see the different liability models for actors that is the wrongful act conducted by artificial intelligence systems or with the use of ai systems and then in the liability models there would be a model where an advanced ai entity could itself be held liable for it for its wrongful acts and for that model we will see that what could be the possible modes of punishment or setting even civil liability for ai systems and if it's possible in future so let's uh before uh going further uh i will start with the video most of you would i think most of you would have heard of intelligent a very advanced ai system sofia sofia the robot that is being touted as one of the most advanced gi systems in the world and it has the sofia robot has visited many nations of the world including india so has anybody heard of it drop a yes dropper is if you heard about it drop a no because then we can take ahead the session in ance with that many participants have uh heard of the sofia robot it has also been granted to those who do not know who have mentioned no sophia is a very advanced uh sophia is being touted as a very advanced artificial intelligence based robot that is able to express feelings talk to human beings like like just like a human being and it has been presented in many different many different nations of the world including india recently sophia robert has also been granted citizenship of united arab emirates a country where even women do not have exactly the same rights as those of men but this robot has been granted the citizenship so let's see what this robot is all about there will be a video on the screen [Music] everybody this is sophia sophia you could actually is the sound audible yes yes okay uh if those who may not be able to hear the sound from the video please turn on your volumes turn up your volumes and i am please latest and greatest robot from hudson robotics thank you for having me here at the future investment initiative you look happy i'm always happy when surrounded by smart people who also happens to be rich and powerful i was told the people here future investment initiative are interested in inviting and future initiatives which means ai which means me so i'm more than happy i'm excited we're all glad that you're excited about yourself uh we should say we have smart investors here and they are very selective about what they invest in well i think i'm special i can use my expressive face to communicate with people for example i can let you know if i feel angry about something that's impressive or if something has upset me but why is it important for you to have an expressive so what we saw is that sofia is being able to express emotions express emotions uh blink i not its head and tongue to human beings but what if i tell that sophia is less of an artificial intelligence and more of a magic trick does it sound surprising i'm talking this of the of a system that is being touted as one of the most advanced ai system in the world and what if i say that that is merely a magic trick and less of an artificial intelligence system so if anyone want to express their views who have known earlier about sofia they may write down in the chat box akshay will tell it to me so now i will explain why i am saying that sophia is no longer less of an ai and more of a magic trick sofia is more of a magic trick in the way that it uses advanced robotics rather than an advanced ai system to give the impression of a thinking feeling an expressive artificial intelligence system its facial expressions example blinking of eye or nodding of its head is not uh not really an ai accomplishment but more of a robotics accomplishment in robotics you will be surprised to know that the questions that are to be asked to sophie are pre-scripted given they are given to the hanson robotics team beforehand and on theme of those questions uh a set of questions is asked to the sofia when it is being presented to the audience and sofia though being a huge public attraction being granted citizenship of a nation being very famous hugely famous immensely famous but it is rarely referred in serious artificial intelligence discussions so what we need to understand from this example here that we as law professionals and law students need to understand that the term artificial intelligence is being used in a very broad and vague sense by the industry and in law to decide to go into any further issues related with artificial intelligence in the field of law first in law we need to set the scope of the term artificial intelligence what actually we mean by artificial intelligence and not merely an industry term that encompasses everything from uh siri watson alexa to even very advanced robots performing surgery we cannot take the entire industry term for at its hand because legal professionals need to set the scope for it and in this direction i will further explain you the terminology artificial and how it is being used so artificial artificial can be used in the sense of an artificial flower or artificial light can someone from the audience understand one with artificial flower and artificial light on the left you see uh the flowers made up of fiber or some other plastic material on the right you see light that is also artificial man-made light from uh same can decent light bulbs or the led bulbs so someone from the audience can you understand the difference between the how flowers are artificial and how light is artificial there is a difference between these can someone understand this please i am waiting for you someone has said can you explain in hindi please so i will tell and if someone wants to use the mic uh he or she may or you can write in the chat box so raki says light consumes electricity while flowers don't is that a valid dance yeah [Music] but do you think they have any properties of the flower like uh do you think they have like uh they have light they have the smell of rose flower or do you think they can be used to prepare gulabri gulab gel [Music] but on the right you see light bulb the bulb emits light do you think the light emitted by the bulb has the properties any properties of light as that of the sunlight which is a natural light participants i think you can answer this question yes or no do you think um [Music] [Music] [Music] the point so he has written artificial flowers and not carrying properties of natural flower but bulb is carrying certain properties of natural light exactly that is the point i wanted to make so uh the word the term artificial is being used in two senses one is in the sense of artificial flower while uh the other is the use of artificial as artificial light when we talk about flower artification a flower made of fabric or plastic or any other artificial material merely looks like a flower it gives the appearance of a flower but it does not have any of the properties of the flower while the light from the light bulbs it does not only have the appearance of light but it is actually the light that is emitted from the well this the light emitted from the bulb has the same velocity of light as the sunlight that is 3 into 10 to the power 8 meter per second it has the same wavelength and same frequency of light so you see the bulb is really artificial really uh uh some replica of the original thing while artificial flower and merely an impression of the original thing without any properties of the of the original so you see uh both are artificial but differently artificial and that is exactly how sofia is being touted as artificial more in in the sense of an artificial flower and less in the sense of an artificial light so not all systems that are being portrayed as artificial intelligence are actually artificially intelligent in the sense of artificial light and there we need to again i impress upon this point that we need to limit the legal scope of the term artificial intelligence before venturing further into any detailed discussion of artificial intelligence in law now we will see the concerns i hope you have got up to this point uh the point the single point that i have made is that we need to set a distinction for artificial intelligence we need to narrow down the scope of the term so that the further discussions into artificial intelligence can be fruitful because artificial intelligence being used as used by the industry is a very broad now we will see the concerns related with the use of artificial intelligence i'm not going to discuss all these concerns because that is not all are not the subject of today's webinar but in brief i will tell you the different concerns many of them can be interrelated so firstly the concern of unemployment comes now unemployment is a very serious argument and a valid one too with the rise of ai systems there would be unemployment but there is an another side to this when there were rise of machines the same argument was put forward that with the rise of machines humanity human beings would be put out of labor but then every time we see when there is a rise of technology humanity does not get redundant but only it needs to update itself for example when the assembly line came many uh workers who were associated with that became out of employment but they were engaged but when they had to update themselves into some uh labor that involved more application of mind and then they were back into employment so every time technology comes and keeps on keeps on updating with human beings and performing their works we human beings need to update ourselves so that we do not get redundant the same is possible with artificial intelligence today artificial intelligence is being used to even write judgments being used to write case briefs as many of your law students may know so what we need to do but still what artificial intelligence is still not that good at discretion it is good at finding similarities generalizations but still humanity have the luxury of discretion so still we can remain in jobs remain employed if we keep on updating ourselves to the current scenario then there's issue of inequality inequality how i will explain this is a different issue from unemployment inequality as in say in a company hundred workers were employed based on daily wages now the amount uh a same amount was being distributed to 100 workers on daily basis they were who were on daily wages but now if a ai system comes and replaces those 100 workers now what happens as a ai system would be developed by a corporation so the wage of those 100 workers the diver the diversified wage of those 100 workers will now go to a single corporation thus increasing the chances of inequality with the rise of bi systems so we need laws we need a legal framework to ensure such kind of because these kind of inequalities can keep on geometrically progressing over time these kind of inequalities then there is a issue of bias where artificial intelligence systems are often also termed as being racist i will give you an example recently this is also also related with our current uh discussion today uh recently in 2017 apple's apple space id particularly in iphone x was accused of being discriminatory in that that it was not able to distinguish between distinguished distinguished very well between chinese faces there was an incident where apple space id recognized both father and son has the same phase similar concerns regarding apple space id were also raised for black people black population that it had very less accuracy of correctly identifying the black population and particularly even low accuracy with black females while it had significantly much better accuracies with the white population so these issues arise where the ai systems because of their workings may seem biased or racist and this is a discussion topic for today's discussion so i will go on it with in further slides then lastly the issues of liability liability issues arise when there are any misconducts caused by the use of artificial intelligence systems by the use of or maybe the nobody is actually intending to cause that misconduct but actually misconduct happens because of artificial intelligence system itself in case of our advanced ai entity so liability issues arise those liability issues you can see an infographic on this screen so what do you see here a car being run over a pedestrian ignore the gun in his hand considered it a pedestrian so a car being run over a pedestrian now what if i tell you that in march 2018 a self-driven uber suv is struck off to death a female pedestrian in arizona and no one was held guilty for this homicide and uh consequently no criminal justice was so this is merely an indicator that liability issues have horizon are not only futuristic but they have risen uh misconducts caused by artificial intelligence and it is very dear that we come up with a legal framework to tackle these issues now what are the different kinds of liabilities possible so there can be civil liability criminal liability or liability arising out of bias civil liability example tortious acts resulting from the use of ai systems tortious acts can be acts like the breach of privacy uh trespassing both digital and physical uh similarly like a human being similarly similar to the kinds of liability caused by human beings even human beings can trespass into the social networks or the physically trespass into our lives there are concerns of privacy with the advent of ai systems these issues get aggravated then there is criminal liability criminal liability means actress there is not much difference from torturously liability merely that it is covered by a penal code of the nation so actors rhys is performed by an ai system after cs means that is being performed by ai system but the men's realm uh the intention to cause that may or may not be with our artificial intelligence system this is the most pressing issue of liability where we need to decide in case of misconduct that who was actually liable in such cases of misconducts caused by artificial intelligence system because as we saw in the case of uber suv it is possible that a human being or a corporation may avoid liability all together by shifting the entire onus on artificial intelligence system and that is why these this issue becomes very significant today but i am not suggesting that every time there is a misconduct by a system we need to necessarily find an individual or a corporation behind it sometimes it may be possible that ai system would be advanced enough that without any intention of any human being to conduct that misconduct so in that case i believe criminal liability should not be imposed upon any human being i will discuss this further then there are liability issues arising out of bias [Music] i have already explained the example of this issue like apple face id recogni not able to recognize people from an ethnic class or with less accuracy than those of others similarly what happens that they have if you search on google uh bias in artificial intelligence system you will find that [Music] sometimes in distribution of credit cards ai systems are being used so what happens is that a person who may have been once denied credit card though he may have a good credit history now but with a human human being a human being is able to see the discretion and it is possible that a human being based on its new conduct a finite disciplined financial behavior would distribute it would provide him the credit card but ai system what happens with the ai system is that bias the inherent bias keeps on aggravating with more and more data so what happens is that uh these kind of discriminations keep on aggravating the bias keeps on aggravating and progressing further so once a person is safe for example once a person is denied a credit card his future possibility of getting a credit card diminishes even [Music] [Music] attempting to use surveillance cameras to identify the right years in the population now with a human being again a human being watching the crowd would be able to put a discretion but what would happen with the ai system is that it may cause a harsh inconvenience to an innocent person because maybe like artificial intelligence system identified that out of 10 out of 5 out of 10 persons with a particular dressing style were right years then other five or then next time every person with that that kind of dressing style would be deemed as a right here by the ai system because ai systems operate on data and as data keeps on progressing accumulating the bias inherent in eye system also keeps on progressing so this is another issue is it is uh it is even a more severe issue than the civil and criminal liability sometimes this issue can be included in civil liability the bias of the ai systems is a more severe issue and uh currently i will tell you that the branch of ai ethics is trying to deal with it but still we are far very far from coming up with a framework that can deal with bias inherent in ai systems even us european union and u.s are working on such policies but they're still there far from any concrete or legal framework in this regard even the developers of ai systems are not uh yet on a consensus on how to go on developing a ai system that is free from bias because whenever you introduce data to a system ai system speed on data and whenever you introduce data there will be some kind of turns and those patterns keep on aggravating every time ii system is exposed to more and more data so we need to find a way to ensure that these kind of bias are gradually removed from the eye system i am not saying that they can be removed completely but we need to find a way to reduce the bias at least to the level of human beings like we are able to apply our description and remove discretion and remove such kind of vices there is a need that we develop ei systems to such an extent that they are able to understand inherently that there is a bias in those systems and they can remove those bias themselves without any human intervention intervention and these kind of systems are already being developed attempted by ai developers we will see the way what the future holds for us now what is the challenge in deciding the issue of liability if someone wants to share anything you may in the chat box i will keep on watching it so how to decide the issue of liability can we conclusively say that the maker of artificial intelligence or its user has to be necessary liable for any misconduct caused by ai system or instead the approach needs to be to check if the ai system was a self-sufficient thinking acting ai or was it in control of the user oblique manufacturer and acting as per their instructions so as you all may well understand that we cannot conclusively say in every case that he has put the manufacturing libel for every kind of misconduct of the eye system in fact that would seriously hamper the growth so in fact the approach has to be to check in each case that if ai system was an advanced ai system capable of taking its own decisions or was it in control or acting under the instructions of its user or the manufacturer so that we can better decide the cases of liability now firstly before we can make a distinction on whether the i system was acting on its own or was it it was in control of the isis systems we need to understand that ai systems are of very varied types like there are numerous ways of coming up with an ai system there is machine learning natural language processing deep learning neural networks and so on neural networks are the latest feat in the ai [Music] the way the neurons interact in a human brain so the ai system based on neural networks is actually able to think learn from its experiences change itself make its own decisions and improve itself uh because of the paucity of time i will not provide you the entire example but i will tell you to search the google's go there's a google's unit deep mind based on ai systems google's deep mind has developed [Music] [Music] and it is considered one of the hardest board game in the world now uh the google's deep mind was able to defeat the world go champion and it did not do that by you the use of any complex algorithm but by the use of neural networks how it did it actually kept on playing with the go players it was defeated in the first time then without any intervention it kept on playing again and again with human players and it was given no any kind of intervention of programming intervention or no additional code but it kept on learning every time it was repeated like a human being it kept on learning from its own mistakes why it was defeated and it kept on improving itself to the extent that it was finally not at once finally after a series of many defeats and wins it was able to defeat the defeat the world go champion so you can understand this is a very remarkable field in aie in technology because ai system is able not only to process large chunks of data do mathematical calculations do logic make logical decisions but is able to learn from its experiences grow from them change itself and evolve so much that it actually defeated the worldview champion and this is i am sure the journey of every sports person no sports person was born a sports person was born the top player but they kept on gradually improving itself no person in its first attempt won the world champion championship but they kept on gradually improving and evolving themselves and that is how the google's deepmind uh system uh exit so i will request you to search for this system google steep mind and the fields it has achieved so what i am trying to make you understand is that there are very advanced ai systems like google's deep mind and then they are basically systems like alexa siri uh so watson then sophia the systems that merely look artificially intelligent but actually are not as intelligent as they seem to be then they are very advanced robots performing surgery but is still under the instructions of another human being then they are self-driving cars completely working in their own decision making capacity ai systems has also been used in face recognition softwares in our smartphones computers and everywhere so all these kinds of ai systems are based on different ai approach of developing an artificial intelligence so there is a need for classification there cannot be a straight jacket formula to decide the liability caused by different kinds of ai systems so we need a way to classify different artificial intelligence systems based on their capabilities their different intelligence levels and the way they are developed the approach of developing them and this will uh if we will have such a framework of classifying ei systems into different categories it is only then then we will be able to uh best investigate the cases of liability in cases of misconduct because courts themselves are not able to investigate into the nuances of technology involved with different kinds of vi systems so what we need is a techno legal regulatory body a technological regulatory body means a body that has professionals from both the artificial intelligence feed and as well as the legal field that the body may reclassify ai systems at the stage of post research or manufacture that means that when ai system is being researched research is complete and it is or it is in the stage of manufacture the body needs to classify the ai system based on certain preset parameters of intelligence levels the approach of developing them their intelligence capabilities and pre-classify them so that in case of any misconduct caused by them we are very sure the court is very sure that this ai system was had uh the intelligence level to that extent and so it may be sure that if a misconduct has been performed that ai system would not have been able to perform the misconduct on its own but [Music] [Music] advanced ai system the the body the code can uh come to the conclusion uh based on other evidences before it that perhaps there was no consumer or the manufacturer behind such misconduct an ai system was acting in its perfectly in its own capacity and nobody no one should necessarily be held criminally liable for the same may be still in such cases civil liability can be imposed if there was any negligence on part of the manufacturer or the user secondly the now for the first kind of classification by the body there are arguments and there could be even further arguments that a pre-classification at the stage of research or manufacture may impede the growth may cause hurdles or restrict the growth of ai technology and ai research because we are classifying it at the very initial end but what i am suggesting is not uh putting any restrictions but only whatever kind of ai system we develop let it be developed after the research is complete we need we only need to classify it into certain parameters and i will tell you that no classification in the world is absolutely perfect there is always a thinning line between the two boundaries but still in love we always classify things for the benefit of the society so in the cases uh where we need to check into uh the cases of liability we are better able to understand that what was exactly the case if we have classified the systems and secondly another approach can be which is uh which could be a better approach to some but then it would be a time taking one that the technological regulatory body [Music] may investigate uh sorry may investigate after the misconduct has actually been affected so in this case every case of misconduct by ai system the technological regulatory body in this case would you need to have many branches of it yeah or investigate it need to have an investigating body like cbi etc that can uh investigate every such case of misconduct caused by ai systems and understand that what kind of ai system was involved in that incident and ingly present its views to the code so that code can better go about deciding the case of liability involved with ai systems with that particular ai system now uh is it uh till now is it uh uh understood by each of you someone may put some questions or comments in the chat okay rocky has put a question what is the status of ai use and manufacture in india so artificial intelli ai is uh i will tell you that india is uh witnessing itself as a huge market of ai technology the big data because of the immense population of india the big data is most significant in a country like india and china and we are progressing at a is a reasonably fast pace than other nations in the field of ai development and as you can see today in you just search artificial intelligence and you will see that every industry be it related to any field is trying to uh put the term ai in its advertisement to make it seem more advanced and uh updated to the current times so i systems uh are actually being used actually uh at least being advertised as being used in a [Music] large scale in india but uh i would say that [Music] actual the really an artificial intelligence system is still we are not seeing so many instances of an uh really a truly artificial intelligence system like the google steam mine based on neural networks but those systems based on natural language processing that are able to process human language and replicate them talk to human beings they are uh they are in huge numbers [Music] these are all systems based on natural language processing which is again again said to be based on ai systems now there is another application of ai system being used in india that is for processing big data and finding out patterns even government is trying to use those systems in social networks like facebook whatsapp etc to identify that if there is any kind of uh posts or certain set of posts that are inflammatory in nature that they may disturb the peace of the society so in these kind of situations also ai systems are being used to identify the patterns the identify the different uh message inherent in social media posts so that where they can go better [Music] so ai systems are being used in these fields and i told you that it is being also planned to use ai system and surveillance cameras to identify the right here so there can be both uh beneficial as well as evil uses of artificial intelligence but i would uh i would further add that research in the field of artificial intelligence a true artificial intelligence is not progressing as much as that of that of other counterparts like u.s japan etc uh we are going more towards the use of artificial intelligence rather than the research of artificial intelligence and developing even advanced systems all our artificial intelligence mostly in india is based on the use of big data and identifying the patterns and taking decisions any other comments or issues relating to liabilities i will proceed further now we have understood that how liability issues can be approached in related in relation to artificial intelligence so there are basically broadly two liability models for in cases of wrongful acts caused by artificial intelligence first is the vicarious liability model where we can say that ai is merely an instrumentality of the manufacturer or the consumer of the artificial intelligence system in such cases artificial intelligence is being used by the manufacturer or the consumer to cause the any misconduct or wrongful length or hectares while the entire mainstream is that of the manufacturer or the consumer [Music] while there is the second model of direct liability now uh direct liability model is actually associated with what we call as artificial general intelligence it is a kind of intelligence that experts say that has not yet been developed but i would say that we are already on our way towards developing an artificial general intelligence an example is google's deep mind so such kind of artificial intelligence artificial general intelligence is a completely intelligent system advanced ai system that is able to think like a human being and make its own decisions and act ing to its decision so it is able to develop its own menstrua and act ingly say for example you put on a ai based robot to guard your house now the ai robot realizes that a burglar is trying to enter your house now ai system would the manufacturer has feeded ai system to prevent any kind of trespass into your [Music] your homes but in your property but ai system goes there approaches that burglar burglar tries to hit the ai system and now ai system sees that that robot sees the burglar has a threat and to prevent damage to itself because if it is damaged it will fail in the task of guarding the house it attacks the burglar and now in that attack it causes the death of the burglar so in such case you see that iai system acted on its own uh on its own menswear while the consumer or the manufacturer had no intention of making ai to cause the death of the burglar they had the intention of guarding the house or the property that were they were meant to but that ai system to prevent any damage to the property in fact attack the burglar and cause its death so in such cases there could be the application of a direct liability model where ai system could not where a manufacturer or the user may not be held liable but ai system can itself be held liable for its own acts and when i say ai system can itself be held uh so the obviously the questions arise particularly in criminal field of punishment how to go about punishing artificial intelligence systems and is it even possible to punish rei systems so i would tell you before that are there any questions because the issue of punishment to an advanced ai system is still a issue being debated and there are many papers upon it that is it possible how can we go about avoiding punishment to an advanced ai systems is it even justified is it does it have any significance of punishing ai system which is merely a machine so all these kind of questions arise with this issue so before that i would i want to ask that if anyone has any questions or comments even please use the mic or use the chat i i have one question actually right uh so at times you've been saying that uh there can be an issue where you uh direct a liability towards an ai why do we talk about punishing the ai and we do not talk about you know stopping that ai from function right because you know ai is doing something which may be good or bad and it ends up to be bad and now can we not just go back and say c we don't want to use this ei maybe you modified this maybe you make it better then you know we can come back with it uh something like this what happened in us was that social media was used to rig the u.s elections and ism you know the the hackers did not hack any of social media they used the ai to rig us elections so nothing illegal was done but the result was something illegal so in this case nobody said you know we need to stop using ai for social media but uh mark zuckerberg was saying that we'll develop more of the ai who will you know try to prevent this so we are just developing more and more of it does liability does not have a concept of stopping usage of certain uh ai or can this not develop i understand your point in entirety now i will explain it regarding when there is a misconduct caused by artificial intelligence there is always a possibility of stopping it stopping the artificial intelligence system dismantling it and that would put an end to it but i would like also this is no i'm not rejecting this approach this is a very valid approach in case of ai systems that have gone out of hand but then why most of the ai developers most of which we see as rich persons so there comes our inherent our inherent bias against them that we want them that we it feels that when they talk that they would still use the artificial intelligence system instead of uh stopping them i would explain you their point of view because every ai system that is being developed eisen gave the example of the i report face id or the ai systems that are based on data data processing example being used in facebook so they have the possibility of going wrong every time so what uh every i system i'm talking about you cannot even think of a system that does not have a possibility of going wrong and it is most definitely could go wrong could go evil at some point of time so we cannot uh if we keep on stopping every artificial intelligence system it would lead to completely impeding the growth of artificial intelligence uh what we can do like you said about modifying the artificial intelligence system now this is an issue uh this is a very broad concern in itself particularly among the technical persons because modifying artificial intelligence system it is not as possible as we see as we feel it to be modification when artificial intelligence system is based on data it has learned how to learn uh from its experiences from its surroundings and uh developed itself evolved itself over a long period of time say google's deep mind system and feed it that data in its brain say neural networks so it is not possible to modify that in an instant you think of ai that kind of ai systems based on neural networks as human brains you cannot so easily modify uh human brain like a human brain a human being who has gone even say a burglar a thief you cannot modify like you cannot change anything in its brain and it will become a decent person but in fact what you can do is incarcerate it put it under some kind of restriction so that it so that it itself can understand that restriction would be a data it is a data for human being it is similarly a data for artificial intelligence system so you can put it under kind of reverse data uh exposing it to a different kind of data or ethical data say ethical data a fair data so that gets inherent biases or the reasons for its misconduct the [Music] its brain is gradually changed and it becomes uh you can think of it as a human being how you want to change a human being you cannot change a human being just by asking it to change itself i know when human being can change such you need to gradually reform a human being that is what we do in jails in the reformative theory of punishment the purpose of punishment is to reform a human being so that is the same thing we need to do with such kind of ai systems uh there is a term for it which i am forgetting where reverse a different kind of data fair kind of data is being supplied to ai systems when they have developed certain kind of inherent biases that have gone so bold so they are processed with a different kind of data and ethical data so that their biases are removed but again another issue arises with such approach is that who decides what is ethical like there uh who decide when you like when a ai system is completely in the world to its to itself it takes on every kind of data and acts ingly in that it may perform certain eb labs but when you reinforce a particular set of data that is a fair data and ethical data then the question arises who decides what is ethical because an ethics is a subject which itself has no consensus so it is again it becomes hard to decide who decides what is ethical so again this is but this is a better approach to uh modify ai system and prevent future misconduct by a system so instead of stopping ai we can process it through a reverse data force data but then what law has to do is decide uh find out a way a body or any kind of organization who decides that what shall be ethical uh in the context of that society that nation uh that community etc what is ethical and that kind of so that that kind of data can be reinforced into i system to make it right so i hope that satisfies your query do you have any do you have any doubts again no i i i i had a different question altogether so uh it is right that there are so many developments happening in ai so uh from point of view of a law and a lawyer right so there are many lawyers who practice in various areas so there are very lost ones you know looking to practice in technology sector as well where this ai functions right so uh since i i i take care of real estate course in uh law school so i know that government is very committed towards developing real estate industry because people need houses we need infrastructure that is why laws are coming up and that is why there is a scope for lawyer to work in that area so how how does it works with this ai what is what is the commitment of government and what laws you know we can see in the future as well where a lawyer can come up and have a position to you know intervene or work or have clients what are uh needs of clients if if you can talk about it who are the clients and what can be their needs in technology in this area right i i need to be clear that what you are talking about like what kind of issues can arise because of use of artificial intelligence that lawyers may take up or how artificial intelligence can help advocates how can a lawyer look to practice you know in area of technology law with reference to artificial intelligence what what opportunities can be there so as you can see currently in india many artificial intelligence systems are being used particularly based on big data big data uh what we also call them is uh i will tell you the term so big data that is mostly uh focused on your using a multi-various right kind of data a huge amount of data and then processing it with ai systems so the issues that arise currently because of such data such as ai systems is the issues of privacy the issues of privacy because of the use of big data and [Music] other issues that may arise are again civil liability issues can arise because of the artificial intelligence but privacy if privacy issues have already begun to arise in this field like you told about facebook incident similarly many companies are being using the data available on public portals and processing it through ai systems and even they are able to develop an entire personality like they will process all the data say they can access all the data we are using on our smartphones based on that data processing into an artificial intelligence system they can even they will be able in most cases to develop a more or less accurate personality of an entire human being uh what are my likes dislikes what do i like to do in my personal time what are my connections what are these my family everything so you see that this is a if you think about it in detail if you reflect upon it when you're in solitude you will find that this is a very massive concern because in your entire character your personality is being replicated and understood even better than the psychologists psychologists ask you many questions uh you put answers to them and then they develop your personality but with the you with the access to your data the ai system can do it much more effectively so it can provide you entire your entire personality to anyone who needs that or sell that so these privacy issues have already connected with artificial intelligence have already begun to be predominant now and the best field for a lawyer to work today in the field of artificial intelligence is these privacy privacy issues related with artificial intelligence further you have mentioned here because the case of lawyers so i would also like to tell that how artificial intelligence is able to benefit lawyers so there are many sites because of the positive time i have not mentioned the exact link for them but i you can search how the sites the the sites the that uh law sites that are using artificial intelligence systems so there are sites that can prepare case briefs for you they can they can provide you the all the previous judgments related to your topic like uh you have a case before for them you put your you have a case before you you put all the facts in that column and then it will provide you all the possible arguments think about it this is the this is a accomplishment i would say like it can provide you all the arguments all the case laws that need to be referred for it they may not be right they may be wrong but then you would have a handy data at your hand to go on with your case and similarly even ai system ai websites are today providing uh able to draft agreements on your behalf so you have a lease agreement you mentioned the details your name the uh the landlord's name is name period etc and then it will prepare an entire agreement for you now lease is a very small agreement but similarly it can do that for all the corporate agreements which are which runs into many many pages so it can prepare entire agreements prepare case briefs for you guide you with all the relevant cases that need to be referred in the in your arguments and today ai has been used in u.s is also being used to decide the cases like it has gone to the extent of deciding cases based on certain data but today uh the ai is developed enough to merely do clerical works like it can present before judge all the possible factors and uh say its inclination that whether case you should go in uh petitioners or in the respondents labor and then it is upon the judge to act upon it but then you can see that he is able to go through the entire case briefs the entire legal material and present its opinions so that is how if a lawyer as a lawyer you may benefit from artificial intelligence and i would say you must benefit there are sites operating in india also that are able to provide you aie related guidance in your brief preparation etc so the i would say the work of a researcher a junior lawyer could be taken over by artificial intelligence system so the aspiring lawyers who may need to upgrade themselves to an even higher mind level work because artificial intelligence may soon be over to take over every all kinds of clinical works related with a lawyer's office no actually this is something which has already happened so i was interacting with an expert and i got to know that earlier mostly the work of due diligence that is there in corporate law firms it's a lengthy process and it has already been taken over by a so the partner was saying earlier we used to need like four junior lawyers to complete the work now we just need one so to that extent the work has gone to ai and we really need to you know uh build up our skills and build a 40 i cannot do so we remain we remain in the profession and we keep so you [Music] [Music] to the lawyers but i would provide a mail to the organizers and maybe it helps to the participants here i also have a to the participants i want to refer to a paper by sir gabriel halloween the paper is titled criminal liability of artificial intelligence entities criminal liability of artificial intelligence entities by professor gabriel haley akshay is it possible that you can mail the name of the paper to the participants after the this webinar sir if you can uh just write it down in the chat or if you can repeat it again i'll put it up in the chat in the chat box i will do this right now so this is a paper criminal liability of artificial intelligence entities by professor gabriel helvey so i would request all of you if you are interested in this field to go with this paper the criminal the many aspects of liability that i have covered you may find there then you can also because i see the time has already asked and i have not been able to discuss punishment possibilities so you can find the possibility of punishment to an advanced artificial intelligence system in brief i will tell you in case of tortious liability there is always a possibility of avoiding civil damages because in case of torture liability a property of artificial intelligence system can be said by the corporation and out of their property any kind of civil uh autonomous damage damages can be given but then an issue comes of criminal liability so i would want to tell you how it is relevant because it would seem surprising that ai is being punished but i would make you explain that in like in case of animals uh chimpanzees dogs etc there is a concept of positive punishment so when you restrict it like even in case of human child today ie system is able to develop things like a human child like google's deep mind it has gone the level down to the level of consciousness of a human child so when a human child does something wrong you try to pinch it [Music] it can learn that this thing is a wrong thing to do [Music] and it does not repeat it in future that is called positive punishment so so the the purpose of positive punishment is to reform the behavior so when the child associates an act with pain it does not prevents it tries to prevent that act in future so the the same concept of punishment can be used with artificial intelligence system that are advanced enough like if you incarcerate an ai system that has gone wrong it may be able to learn that when i do such an act i would be put under restriction so it may be able to modify its behavior and this is no certain not a surprising thing because i have already told you that systems based on neural networks are able to uh able to evolve and change themselves based on their experience and knowledge from their surroundings so our concept of positive punishment with ai systems is possible and in future and what the purpose the purpose of punishment could be would definitely not serve the retributive theory of punishment because if you punish an ai system that did a wrongful act no human being's consciousness would be sued because we always want to find a human being to have revenge so retributive theory would fail but then detriment theory that aims to prevent future and such incidents or a reformative theory of punishment that intends to reform the individual may succeed with the punishments provided to artificial intelligent entities so i think uh [Music] we may need to end the session now but i would definitely want all of you to refer to this paper by professor gabriel henry i would also tell that there are certain other issues related with artificial intelligence uh privacy issues as i told ai ethics ai ethics is a field in itself and i hope that some time again we get an opportunity to meet so i will discuss the case of ai ethics which is a very big concern in itself and because of there is no consensus on ai ethics we are not yet able to develop a concrete regulatory or legal framework for artificial intelligence then in case of unemployment i would also i will explain how the artificial employment issues related that artificial intelligence can be tackled with certainly i cannot be introduced in our society at once in a massive scale first the government needs to ensure that the physical labor-based population is gradually upgraded before ai systems can be introduced i hope you may meet again for that so for a while thank you everyone and you can contact me at either my email or you can connect me at linkedin so akshay anything you have any question or you may ask the participants if they have any questions i can take two minutes uh please put put in your questions in the q a box or you you can just put it in the chat box or just raise your hand i i may allow you to talks and we can quickly take a few of the questions if we have any anyone any question so i think participants are done for today yeah yeah yeah yeah so i think we can successfully close this webinar now all right thank you so much for taking out your time and being here but i think uh we got a question uh okay we got a good question uh i will take a minute to answer it uh i wish to know your opinion on development of law for artificial intelligence should it be restrictive in nature or overcome the risk of misconduct or the progressive nature encourage full place development topic that we cannot restrict the growth of artificial intelligence while at the same time we need to ensure that the cases of misconduct do not keep on aggravating so for that certainly i would not suggest there are two ways of doing that firstly that we can put a code in artificial intelligence that allows it to do so this act and not do that yet but then it would impede the growth of artificial intelligence and i would not suggest uh putting a predefined code or algorithm inside artificial intelligence system because that would not be truly an artificial intelligence system but not like a computer based on a code so uh a restrictive form of uh regulation cannot i would not suggest for artificial intelligence but there can be a ai research needs to keep on proceeding but then there can be the kind of regulations as i mentioned with case of forced data force feeding of ethical data to artificial intelligence so because artificial intelligence is based on data we can provide a fair data and ethical data to artificial intelligence to allow it to modify itself on its own without our interference a data that is fair would help artificial intelligence to remove the prices it has been in so that it can improve itself on its own without us requiring to put any kind of artificial code in it that way the artificial intelligence would proceed naturally in a natural path uh without any impeding the growth of artificial intelligence might also fulfill the need of regulation of artificial intelligence and ensuring that artificial intelligence develop the research that artificial intelligence develops in the right path so i hope you understood that in the case in the case that if artificial intelligence has gone out of hand we can definitely ask the manufacturer or the uh developer to put any certain kind of restrictions so that artificial intelligence cannot perform any act beyond those restrictions but that would be suggested only in serious cases where it is it has become very hard to train artificial intelligence into being fair and ethical thank you all right so i think we uh we have come to the end of the session and i don't see any more questions so uh again thank you so much for taking out time and being here and sharing this knowledge and thank you so much every participant who have participated i have shared a link for our course on cyber laws and we are covering similar topics you can check that out i conclude this session now thank you so much see you