welcome welcome to another episode of the impact of data and AI this week this episode is sponsored by Wilde women leaders and data and artificial intelligence Our Guest today um if you're following her on LinkedIn Aruna betham she is the head of AI and analytics and data science for the asia-pacific region at capgemini welcome hi Melissa I'm glad to be here morning yes so that's important because you're calling in from Sydney New South Wales I'm on the Eastern time zone of the United States so it's 6 p.m here and 10 a.m for you yeah funny my uh my daughter says that you are in the future and I am in your past I mean it can be taken good and bad because I'm in I'll call it down under so uh you know timing with global becomes a challenge all the time but yeah but you have been really busy um in 2022 you were linkedin's top voice and AI um in 2023 which is just now come you have been highlighted as one of the top 100 billion women in Ai and ethics and you just recently within the last year become a group member for the responsible Think Tank a um excuse me correction responsible AI Think Tank yes yes got it thank you yeah she's she what you can't see for those that are watching this via audio podcast is this really big big grin from ear to ear yeah I mean uh I mean you know to be honest uh when I started this journey this is not what I had in mind uh I was I'm very passionate about Ai and data science and um you know while I was working and going and doing this you know a job and coming back I thought there was something missing in me uh you know got a bit monotonous so I thought after all these years of experience I need to give it back to the community so that's what started this um Journey a few years ago and without any expectations I just did what I um passionately believed in doing and uh I think if you just do it what you need to do things will come by itself and that's what happened I'm very glad about it and definitely and I'm very happy and gives me more motivation to do more perfect as long as you've got the 24-hour cycle to do that that's been my problem is I want to do a lot of things I want to give back I'm actually limited by the time the concept exactly I mean that is still a challenge uh I tried to scrape and I wouldn't say it's it's been easy um a little bit you know definitely support from the family my husband and all helps to the cause but I would say it's not it's not easy at all yeah I I asked you to be a guest and you accepted um primarily because I want to combine you know the the combination of um the role you have as head of AI and data science and Asia Pacific but you also have visibility into the rest of the world um but in focusing also in the combination of your top LinkedIn voice and Ai and then also what's going on with the responsible AI think tank and all of that I think you've pulled together just recently in your posting around the top five Trends in AI Technologies I think that's something that we can absolutely unpack for today yeah sure sure you know definitely and thanks for having me here and hello to your audiences uh and um yeah um so as we know now uh AI you know has been there for many many decades and most recently it's been coming into mainstream uh you know a lot of people are talking about it and using it um but one of the things with AI is that it is an evolving field while there are a lot of people uh looking into implementing it it is still evolving day by day right so um so it is important that we keep up to date with the latest and the greatest that happens in the Air Technology and it's actually changing from years to months to weeks I don't know maybe it will go to days where it'll become very hard to keep up with what's happening in AI even um for people like us but um I think that is very important so you know what's out there and how I can use it for you know the benefit of you know business individuals as a society so um if I have to sort of highlight the top five Trends the first would be the AI and cyber security which is a no-brainer I think because cyber security has been such a huge challenge um for everyone across the globe because of the constant threat that we get with the with the cyber attacks that happens uh especially it's more more practical or real for Australia because 2022 has been a really bad year for us um as a community we had two major cyber attacks happen one with Optus which is a you know biggest Telecom in here and another one with medibank which is again a biggest um ielt Insurance um and I see you know the industries that are vulnerable uh in the top industries is a government sector and Health Services because of the rich data that have and and the amount of uh you know money that people can make with those data so uh that is where I think AI can play a very key role in fact I just did a Blog on the data protection day which came about uh last week or a week before on the importance of AI powered cyber security right so um you have the technology that has complicated complex algorithms that can uh track through the you know millions of data points and find out what their relationships are identify threads um and to be able to prioritize and those threads and address them so um AI has a huge role to play uh when it comes to cyber security also also I can also get into bad bad people as well they also can use it for their advantage you know sniffing through looking for uh you know opportunities for weaknesses that can they find in the systems and all the networks and then just immediately automatically you know send something out to to to um so that is why I think it's more important that you know it's how you use it um but either way AI has a role to play in cyber security so that's the first thing so Aruna when we think of cyber security it's identify protect detect respond recover but cyber security is is really the security around the electronic aspects of data information the electronic version of information so then with the discussion of this this first Trend how is AI really helping how is it how is AI working with that cyber security industry and why did you list that as one of your trends yeah so uh you know as I was saying that you know AI um has got complex algorithm it can serve through millions of data points it can see how they are detailed how they are linked together uh UK there is anomaly detection that happen you know what is the normal way behavior of these networks and systems and you know sometimes you know when those things can go up hazard you know anomaly and those can be automatically detected while a few um you know a few years ago we did have uh anomaly detection but that's more rule based right now the ma with AI and machine learning they can learn from the data on the go in real time and find out these anomalies find you know because people tell that's what we also see in fraud uh spices well people tend to the Bad actors there tend to know what you're trying to um you know protect and they go look for a different set of behaviors to bypass that now with the AI it can learn from those behaviors and you know automatically try to prevent them so that's why I think AI has got a much more scope because it can handle the volume it can be uh quite you know speed and uh can instantaneous all of these uh I would say would be a great power in in attacking cyber threats thank you that makes a lot more sense for me yeah I appreciate that so let's move on to number two okay so then the number two I I see is democratization of AR so this comes as AI has become mainstream you know people are starting to understand what AI is what is potential and how it can be applied in different Industries uh what that means it opens up opportunities everyone wants to be part of the bandwagon everybody wants to want it use its potential but there is uh you know challenge in terms of the resources and the skills that is required um to be able to uh you know for everyone to get into it it you know small medium firms you know do they have the um you know funding to do all this so now what we are seeing the trend is um automl no code low code platforms are coming up uh you're looking at citizen data scientists who um you know people who have got the subject matter expert who have been in Industry they know what is what they want but they don't know the technical expertise to be able to use this sophisticated AI systems and that's where uh you know these tools are coming in um uh so which care where you can feed in the data um you know you've got to find out which data is um you know relevant for the problem that you're trying to solve feed in it and automatically looking at the data structure of the data nature of the data I'd come up decide what algorithms to choose and then it comes up with predictions which you can then look at it and provides nice reports and stuff which gives you enough information to have the confidence that it is doing the right thing and then you can review and then make the decision so that means that it has AI systems is now you know coming to everybody's hands and they can use it for the advantage so that is evolving and it I see more and more the tools will be developed um to actually tap into that uh huge Market that is out there who is very keen to use AI so that's my number two yeah I do have one question when it comes to you mentioned the comment about uh resources um so when I think of resources I think of both the individual resources skill sets but I also think of that computing power you know yes you know there are companies like IBM who've just recently pushed out you know faster silicon chips they can sift through the state even faster but we're still finding that's not fast enough so is is that kind of what you meant by resources that kind of wrap in your arms around resources primarily more on the skill sets of the individuals uh of course where you know democratization AI when it comes to Technical Resources we're talking about the power and all that that's probably similar in my view whether you're doing a heavy and data science platform or I think they are try they they are getting more mature in that space I don't see a major difference but it's more primarily on the skill sets you know having um you know thousands of data scientists which is hard to come by uh and all of that you know unicorn uh technical people we are looking to solve world problems so I think democratization will definitely make it much more common and accessible by everyone and two others to just answer your question about the computer power um you know Edge AI I see it's the um third Trend that I see will get more and more um leverage in the future uh because we are seeing uh more and more smart devices and handheld devices being used and we also want um insights and decisions to be made uh when we are using those devices now um you know it'll be very hard if you have to rely on sending all the data back to the servers or have this High processing power and and bandwidth the the good thing about the edge AI technology is that it doesn't require a huge processing powers or the bandwidth it is on the uh device uh the machine learning models runs in the device where um what you're using and for example we are seeing it getting used in medical devices like if you are doing scanning uh you know you could then and there uh you know run some algorithms and detect um you know results uh from those devices rather than you have to wait for I need to send to the server and come back with the reports and with all these smart appliances we are using uh phones we are using um and self-driving cars you know iot anything with iot devices you could use AIG technology to actually uh you know make those uh predictions uh faster quicker and the decisions also can be made into instantly so that is an industry I would see definitely really picking up because uh we are getting more used to handheld devices now and we want Smart Homes everywhere so I can see a huge huge market for that the next one um I would say is generative AI so which is a Hot Topic uh at the moment so so generative AI just to give a bit of a blurb it's either artificial type of artificial intelligence where you give a existing input and it can generate new images new videos new data data sets and so on with that by understanding the features of the input data that you fed in um right so it has got huge um opportunities and potential uh example synthetic data which have been talked about a lot and with the technology getting Advanced you know you could use uh your existing data points and then get new data created and it's very useful when you don't have the right level of data or you don't have the right level of diverse data which is required to build the AI system or if you if you think that those data are very sensitive that you can't use them in in building the AI system so they can become super super important you can use synthetic data don't need to worry about privacy challenges regulations and all of that but still be able to provide the similar outcome because the characteristics of the data points have not changed right it's the value of those data inside has been changed right so so I see um that is one area again uh we are seeing a lot of applications you know you know mushrooming out you know you give a text it comes up with the images you give a text it comes up with the videos which is you know making it more easy for the content marketing um people to be much more productive in in what they're trying to deliver a message and and the last one is chat GPT which um I don't need to say well how why it is important and how it is important it's been soon Army uh or I would say you know nothing has exploded that that quicker as compared to chat GPT again there are you know yes you know it's used in uh marketing and but I can see a lot of applications in business spaces uh using chap GPT technology um right using in customer service while we've been using traditional Bots to actually help with the customer service um but they are much more rules driven um much more not as aesthetic or human-like if I have to say um with this chat GPT um you know because it's been trained with um gigabytes and gigabytes of data and you could use it customize it and use it for customer service where the responses will be very human-like as if you are talking to a human support team person and it can customize the responses based on um you know what the conversation is happening instant stationary so it can take conversational other you know customer service chatbots to the next level um where it doesn't look as if you're talking to some computer trying to solve your problem so that you know call center Analytics where well you are doing a call center you know taking customer calls um behind the scenes it can summarize all the transactions highlight key points identify the sentiments of the person who is uh who's sitting there and requesting those services so um in processing process optimization mortgage processing you know all of that I see a lot of um potential um you know you could use uh digital avatars uh human like uh yeah avatars are coming up with human-like conversation uh you know you could um if you've done correctly uh and if it is Advanced well you know Financial advice you know you can have someone giving some initial financial advisors or if you want to diagnose you know something you have some uh you know um what you can call um you can diagnose the symptoms if you have some symptoms and you wanted to talk initially uh you know it can help in the health system as well so I you know if you keep exploring there's quite a lot of potential in chapter um always you need to be cautious and in fact I put a post today on chapter Advantage but with a word of caution right it's a it's a new evolving technology it's in the hands of everyone which is good and can be dangerous so how do you make sure you use it responsibly so a generative AI is is going to go next level gpt4 I'm told it's going to be 100 volts powerful than gpt3 so we'll have to wait and watch but you know you can see an explosion of advancement in that space I do have one question for clarification the difference between the general the generative Ai and the chat GPD because in in both in both examples it's generating content um so is is the difference because generative AI is more visual within that generative AI you could have generating pictures uh you know Dali is one of the example where you you are able to provide a text and it's able to generate videos and images then you have things like code it can also generate code um for you so that's another type of generating AR then you have this chat GPT which is a conversational chatbot which is generative so they all come under the family of generative AI got it I understand thank you so much yes going to the number five and the most important and passionate topic of mine and quite relevant with the chat gpt's uh thing is responsible AI um you know as I said responsible AI was very important always embed up with chat GPT where it has gone to uh you know everyone even my um my son uh tried chapter GPT I said Mom you do you know there's this charge Deputy it can generate uh text and everything cool and he's trying to be excited to do his uh School assignment with all of that stuff so um you know it's become so so common uh and that is why I think it's important to have responsible AR because there is a tendency uh if not used rightly you can spread fake information misinformation and steer the community towards um you know particular um uh opinion which which might become danger right in the long term so you got to be very mindful of that that's why responsible AI has been a long-term passion topic of me and also that is what we are trying to do as part of the think tank group um in Australia's you know csiro which is our research organization they've got a national AI Center established so one of the objective of the national AI Center is to um you know spread the awareness of AI and responsibly and how do we ensure all the businesses we want to use we want the community to use AI we want Australia to be Forefront in using AI at the same time we want to be very responsible in the use of it so how do you spread awareness of what is the responsible AI what are the tools what are the you know Frameworks that you need to know and you know spreading the awareness educating community and making it much more uh a must to do in the business organization that's what we are trying to do so responsible AI will become even more important and if if it wants if AI needs to be mainstream responsibly I should be by Design and not something that you do later on even before you're thinking about Ari first thing is what is the impact of that AR you know what are the things got shoes that I need to be mindful of so when you design when you do it by Design then things uh can fall in place so that is my last prediction um in fact it's not a prediction it is a must do yes to ensure that we are able to use it more responsibly as we look to wrap up what are the what are the leave behinds you know somebody listening to this says oh great this is good I've got these five Trends okay now what what are some things that maybe they could start doing tomorrow that would help put them in the right mindset so I think uh um you know it's all about awareness and knowledge I guess depending upon uh which industry so that is which industry they are in um uh you know trying to uh well that is what I always you know you need to understand the technology but one of the things I aim to do is understand the technology simplify into business terms and see how it can be applied to your day-to-day problems either in your personal or in the business right that's that's what even prompted me to create this AI in business Series where I talk about how AI could be used in across various Industries I've taken 10 Industries you know banking Insurance retail Telecom government services agriculture Healthcare oil and gas right the potential of applications of AI it's there everywhere and what I would say is if if someone is wanting to use this technology for the advantage age understand which sector you even have a good understanding of um AI generally high level and see how it can be used in your specific industry and and that's where that's what I would um advise them so um some applications are across the customer service across the board in every industry but fine tune your specific you know sector um but you have the technology but you need to sort of customize and fine-tune for your own requirements well I appreciate your time thank you so much for joining our our podcast today thank you so much thanks always pleasure to talk to you
2023-02-21