Prof Das Narayandas Answers The Big Question – Will AI Machines Take Over Mankind?

0:00 given the fact that machines are so much smarter than humans and while everyone's trying their best to come to terms and grapple with the Technologies and learn as much as they can the machines are much smarter so how do you lead and Lead this transformation at a time when these machines have a mind of their own uh actually I don't agree with what you're saying might be because I don't understand uh what you're talking about uh machines

0:26 are smart machines are smart when they fed with valuable information reliable information that they're able to machines do things better once something has been well understood as of today we're still not in a world where machines can figure out things on their own they need an objective function they need content but they can process that faster they can analyze faster but if you if it's garbage in it's going to be garbage out and so as I see

0:58 it the way to think about the world therefore is as I said again its augmented intelligence and think about it as leaders in firms leaders and firms at all levels they make decisions and we are trying to figure out how AI can help them make decisions and if you talk to the world of data science and ask them what kind of decisions are made I mean it's well established these are not my terms but there are three distinct forms

1:24 of decisions that leaders make one is what we call as prescriptive uh Analytics I mean prescriptive analytics informing decisions what is prescriptive analytics you're using data from the past historical data and using machines to process that information and provide you with inputs that the human then uses to make decisions is that going away in the world of business the answer is no because there are still a large number of decisions especially as you go higher up in the organization where data is

1:56 sparse the uncertainty is high and the contingency framework is very complex so if a human has not fed a contingency framework into the machine it is not going to be able to do it unless until it is fed with vast amount of data to do some pattern recognition that's not going to happen so into the foreseeable future all the leaders who are sitting out here I would argue more than half the decisions they make will be what we

2:23 call as prescriptive and in that world you have to ask the question what can a machine do for me the human still makes the decision but what machines will do is analyze the data historical data and present it in a much much better Manner and that's what happens today when you query chat GPD it goes in checks on the body of knowledge and then distills it to a form that you can use to make a decision that's not changing the machine

2:51 does not make the decision there the machine provides you with the background to make a decision the second thing is predictive analy analytics I mean humans using machines to make predictions I mean uh over the last year I've asked CEOs about how many of them have you know a a a a metric called customer churn on their dashboard and virtually every CEO has that customer churn is the number of customers you lost in the last year is

3:21 it a good thing to have absolutely yes because it tells you gives you a high level feel for the health of the business but what do you do with customer churn it's after the fact the customer is already churned what might be better is finding H having a metric which is customers at risk if you have customers at risk the leader can then actually take some action to prevent attrition now machines can do that very well machines can provide in this case

3:50 the prediction and managers make decisions managers still need to make the decision so in Predictive Analytics the augmentation is machines provide a look into the future in terms of predicting predictions are never never good and in the world of business accuracy is not what you want in an AI model because missed opportunities can be very very expensive to lose so what you don't want is accuracy what you want is economically optimal outputs provided by the AI system and then managers make

4:22 decisions and my challenge to every leader in this room and outside is in the coming year make it a point to have at least 20% of your decisions using machines to predict first before you make any decisions that's the way you adapt to the new world and there again machines are not making the decisions machines are predicting and then of course you have the third part which is prescriptive analytics descriptive no prescriptive analytics so descriptive is the first

4:53 one sorry if icri you missed out descrip yeah so it's descriptive analytic the first one predictive is the second one and the third one is know prescriptive analytics in prescriptive analytics yes machines operate autonomously but do humans have a role to play yes I mean I have a Tesla I sit in the car every day once I tell it where to go it does a fantastic job and takes me to wherever I want to go but it still waits for an

5:21 instruction from me and so humans do have to provide the objective function humans do have to provide the rule and then the machines take over but is that going to be important as we you going ahead yes there'll be more and more places where machines can do a better job when will machines do a better job when there are vast amounts of data when there is some pattern that humans might not be able to recognize themselves and there is the opportunity

5:48 to optimize in a manner at high volume that humans cannot do that world is going the machine way but is business therefore going to be run by machines well we're not in that world as yet at least not into the foreseeable future