good morning everyone uh my name is david lee and welcome to today's event on the role of data driven innovation in artificial intelligence or ai as i'll call it save time from now on this is the latest in a series of online events organized by the data driven innovation initiative or ddi if you're not familiar with the ddi it's an innovation network helping organizations tackle challenges for industry and society by doing data right to support edinburgh in its ambition to become the data capital of europe and the ddi is delivered by the university of edinburgh and harry at what university for the edinburgh and south east city region deal find out more online at the exceptionally simple web address of ddi dot ac dot uk data driven innovation and ai have taken on a much higher profile in recent years and today we're looking to join up a few of the dots about how they can fit together most effectively in the rich and complex tech ecosystem which is emerging in the edinburgh city region and the wider scottish economy we've got four excellent speakers to help us with our discussions and i hope you'll have some questions for them and some comments on what they say please pop any uh questions or comments that you have into the q a box and i'll try to get through as many of them as i can during the time available we'll start off with an open question for each panelist to allow them to get some big issues on the table and then move into more detail i'll introduce the panelists one by one as ask the questions again to save time this is a very big topic to get through in a single hour okay so first up we'll go to professor michael rivatsos from the university of edinburgh michael is professor of artificial intelligence deputy vice principal of research ai and director of the base center which is the university's innovation hub for data science and ai so michael a couple of things what what is the role for scottish universities in the development of ai and maybe you can characterize a little bit that relationship between data driven innovation and ai as you see it from from your perspective um thanks thanks for this uh great uh opening question um so scotland was the second place in the world after the us that um started engaging in ai research and this goes back to the early 1960s and i think you know for a relatively small country we're really punching away above our weight in terms of the quality and quantity of the research at our university alone we've got over 400 people working in ai and also lots of people studying its applications or indeed its implications there are several other universities of course with great strengths in specific areas like harriet watt and glasgow aberdeen dundee and so on what i view as the key role of universities in the development of ai is that they are the places where the ideas for new innovations come from in particular when it comes to truly novel technologies and it's only really if you have that technical edge that you can compete with those at the top whether these are the tech giants or the global actors like the us and china i think the other important thing about universities is their reputation and we have universities with outstanding reputation in scotland brings talent into the country and again we can only really complete compete if we manage to attract uh the brightest and the best into the country um the other thing i think that's important is that universities quite often are a trusted and independent partner many people are concerned about the potential risks of ai and i think we have a culture of doing research and innovation responsibly and and this pervades all our activities uh which i think is the strength we need to play to now regarding the relationship to data driven innovation that you asked about obviously not all data-driven driven innovation requires ai but i think what's important is to we're currently in a stage where lots of organizations are starting to learn how to use data to solve their challenges ai is much more about having systems that can help make decisions rather than just look at data and for humans to um um derive insights from them so i think the role of ai in the future will be much more to aid with better decision-making rather than um just having better information and i think uh that will be the next uh level uh of the next step in innovation for our economy okay thanks thanks very much michael some some good points there to get us to get us started off um next up is steph wright who is ai strategy lead at innovation center the data lab and she led on the data labs efforts and support the scottish government in developing scotland's ai strategy and is now working with uh jillian dougherty ceo at the data lab and the scottish ai alliance to deliver the strategy's vision for scotland so plenty to keep you busy there steph so what do what do you think scotland's ai strategy does to help cement what michael just described as scotland's role as a a country that punches above its weight uh in terms of ai um so what will the strategy do and and again where does data kind of fit into that that complex picture um hello hi david and uh hello everyone i have perfectly branded myself here today just to make sure that uh if there's any doubt uh what i'm working on uh anyway yes uh thank you for the question um you know the uh for anyone who doesn't know the strategy was launched uh at the end of march this year and i what it does is it's kind of laid out the vision for um ai in scotland and and that vision is quite simple it is uh that scotland will become a leader in the development and use of trustworthy ethical and inclusive ai uh all contributing to make scotland a fairer greener more prosperous and outward-looking country and uh and so we're signaling to the world that that's what we want to do and you my colleague wanted me to mention the whole you know analogy with cement as an inspiration in your question um you know um the kind of construction analogy there is you know uh was mentioned in our strategies uh welcome uh is that you know this is no standing start we're building on strong foundations uh long history of academic excellence in ai development as michael touched on and a successful and well-connected ai community and a desire to adopt ai for uh more fully across scotland so we're you know we've taken a collaborative approach in the development of the strategy and we uh you know we are planning to continue to take a collaborative approach so we plan to work across everyone you know we're open and transparent process uh to uh realize this vision uh in the strategy we laid out a whole series of actions uh to take over three time frames uh we're currently in the first time frame of first hundred days then we move into year one and year two um obviously you know the the first time frames but establishing say the scottish ai alliance we're currently in the process of um recruiting the leadership circle uh which we plan to have a spread of stakeholders from across all sectors representative of the whole ai ecosystem in scotland and we're also developing we've also begun work on the development and developing the scottish ai playbook which we intend to be an open guide to how we do ai in scotland incorporating our vision our principles and our practices uh so you know the strategy is just the beginning it's a kind of laying out our intentions and our direction of travel but ultimately we want to work with everyone to help us get there okay thank you very much steph for that and i'll go now to helen hastie who is professor of computer science at heritage what university she's an expert in human robot interaction and she's one of the academic leads for the national roboterium the world leading collaboration between harriet ward and edinburgh which is due to open on the harriet what campus in spring next year um so helen how important is sort of data-driven innovation in ai in the development of what will drive the national robaterium thanks david thanks for the questions so the national libertarian is a ddi hub it's a cutting-edge research facility that's as david said being built on the outskirts of edinburgh at harriet watt campus um but it's more than just a building what we're looking at is to work really closely with industry to understand their challenges the problems that they face and co-create solutions um together in order to drive innovation uh forward so that ai and robotics can move you know and benefit society and the economy and net zero so we can all really benefit from ai and robotics research so the national libertarian builds on a work at the edinburgh center of robotics so this is a joint initiative between harriet watt and edinburgh university that was founded in 2014 and the general theme around the research at the center is about safe interaction so as robots and also ai um is starting to be used more it needs robots need to be able to interact safely with um with each other potentially and other systems but also in the environment which can be hugely dynamic um and we have a project called the epsrc orca hub which is joint work with edinburgh and harriet watt and other universities looking at this idea of robotics being able to adapt to dynamic environments and also with humans so robots need to be able to understand humans and we have a large hri component in the center and in the robotarium itself we are having we're going to build a living lab which is basically a flat where we can bring people in and evaluate our methods and our robots um in a in situ in a familiar kind of environment and as we move robots out we need ai ai is fundamental for intelligent robotics vision for sensing um nlp conversational ai all these areas are key so we're i'm very fortunate to have this partnership between edinburgh and herriot watt so we have the research we have the building and what we are looking to also develop is skills and michael mentioned this as well so we have um a center for doctoral training where we're graduating about 150 students over the lifetime of the center and we're training them in key robotics and ai latest research obviously but also responsible research and innovation and also other skills such as innovation and entrepreneur skills so they can go out and you know create their own country and drive the next generation of robotics and ai and finally part of the national roboterum is outreach so we want to focus on engaging young people and the general public raising awareness of robotics and ai and but also increasing inclusion and diversity in the field because we think that's also important as we move forward great thank you thanks very much helen and uh finally i'll come to roy donaldson who is a technical solutions architect at cisco systems um roy where do you think the scottish economy is doing quite well in terms of use of ai and where are the areas where are the sectors with some real untapped potential to do much more ah yeah thank you david yeah so interesting question actually around actually asking about ai i think it's kind of a strange question actually because um let me just frame it in some type of context actually of anybody looking to go and build the house generally doesn't ask their builders what hammers they use and i think that's kind of the context we perhaps actually put ai in you know ai is a is another tool in the kit bag of how you build and you construct actually applications for people how you secure their environment how you transform their infrastructure how you get their teams to work and i think the the whole aspect of which of you know scottish sectors can benefit from that is i would actually actually profess to say actually it's perhaps all across all scottish sectors that can benefit from that i think perhaps you should perhaps you know look more at the things that ai can deliver to people's actual tool sets rather than as ai as a technology itself and at cisco we actually have um ai embedded in many of our applications and our tools for our customers and have had for a number of years already now so ai isn't a a new technology anymore it is something that inherently comes actually as out of the box let me give you a great example um today we're using one of our competitors products actually zoom to actually do this call we have a similar product to market called webex and we recently acquired a company called babel labs and why did we do that because they use ai to do noise reduction when we're on calls like this and you might think noise reduction is simply filtering it's not a dog barking says different from a baby crying and um you know babylog labs actually had researchers at the university of edinburgh you know that's one of the very specific things that we are very interested in them and and that's one of the things we really seek is that you know people don't consume ai people consume collaboration technology that utilizes ai to provide a better experience um to their customers their partners and their suppliers i think it's very much you know we should look at the industry moving forward from actually as a technology to being more as an enabler of the technologies that are providing business value actually to the scottish economy okay thanks thanks roy um just following on from that and touching on what a couple of other speakers have said what about that kind of communication piece how do we get across uh the kind of points that you've just made there roy um that you know ai is a way of you know as you say delivering delivering better outcomes it's part of the tool set if you like how do we get across you know communicate more broadly outside an audience like today about what ai is and how it can actually help support businesses and and kind of support individuals in the way that they live their lives you know because we we we often talk about things within a bit of a closed loop how do we how do we take it outside and really communicate ai in a really clear way yeah dave i think you know really actually to me it's really talking about the business benefits that technology can actually bring to organizations you know really you should be actually looking actually at some challenges within businesses and the solutions that we can actually provide actually to address those challenges for them and i think you know many of those solutions that we actually look to provide actually should actually provide the answers and and we should be able to then clearly demonstrate that there's benefits of using ai to enhance those benefits we can provide for that you know i i you know a great example of that is that um cisco maintains a division called talos we are the largest um non-government funded security research institute actually globally and we we heavily use ai to address security attacks and vulnerabilities and the reason we do that is simply the sheer quantity of data that we receive um we could not do this with without using ai techniques i think that's one of the really you know things we should really be able to actually say to people is that you know if you did this using without using ai you might be x number of days before you could actually look to address a security vulnerability with ai we can actually tell you in 10 seconds what it is where it came from and who's infected and how to address those um challenges i think that's really actually something that's really important we do is i'm saying here are the benefits of utilizing ai to address this business problem rather than discussing the different variants of ai that there are in market and i think the scottish economy and scottish businesses would far more like here actually answers and solutions to the challenges and problems to help them grow the economy rather than discuss technicality details actually of ai okay i'll i'll come to you on that step if i can um just in terms of what roy's just saying about it using you know looking at business problems business solutions um how how do you think the scotland's ai strategy helps to address that to look at those business solutions and and how uh how well connected uh are scottish businesses with the strategy and the understanding of what it can deliver for them yeah i mean you know a very large component of of our work program going ahead with the uh the ai strategy is communication and just engagement with various stakeholder groups including the public uh which was touched on earlier about how you can communicate the benefits of ai but also with businesses um you know that's part of the work we're doing with the ai playbook uh we kind of want to provide you know a kind of guide to a you know either people developing ai or people looking to use ai either in their in their work or you know personal capacity uh so you know it's kind of providing guidance for anyone wanting to get into it part of the another thing we're committed to is around an ai capabilities directory uh but just to uh something i didn't mention in my intro is that you know we're obviously part of a wider digital kind of um strategy in the scottish government so the ai strategy the scottish technology ecosystem review and the digital strategy are kind of a triangle around the kind of digital um ecosystem in scotland so uh you know nothing is standalone uh just like you know date you know ai isn't standalone because ultimately you need data to do ai so when when we kind of talk about the ai strategy we're not just talking about ai so it's about trying to join everything up and kind of collaborating with you know industry academia public sector etc we do have a big focus on the use of ai in the public sector and a big stream work stream around ethical and explainable ai in the public sector uh but yeah i mean the the thing is you know when when we uh during our development process we did a lot talking about yeah you talk about developing ai and stuff but you know not everyone is developing ai you know everyone's making ai lots of people lots of organizations are just wondering how they can uh use ai to improve their their business uh you know whether it's just making small efficiencies or actually completely transforming what they do and so we need to make sure that we cover that entire spectrum uh that anything that includes ai that we can try and provide guidance and ultimately in a way we can look at obviously regulation and ethics and all the much more complicated stuff but in the in the first instance you know we're taking stakeholder input into the ai playbook because we see that as a kind of a reference point for anyone wanting to embark on this journey uh with ai i hope that makes sense yeah absolutely uh michael how well how well joined up are we in in in scotland do you think in terms of all that all the players who need to be part of again doing what roy says in actually delivering tangible tangible benefits from ai even if it is operating beneath the surface and it's not something we we talk about as a thing in itself how well joined up are we and how do you decide from your way you're sitting where sort of research capability is to be deployed when there are so many potential um applications you're on mute michael sorry i think this is this is a very important question i think actually our size and our and our culture and the ways of working we have in scotland are very conducive um to being very you know well well connected and coordinated um and i think we actually can see a vision emerge around how our key sectors and our key strengths can come together so in areas like you know health and care um tourism energy you know the the big areas where we have great data assets we have great um uh industry um expertise and and and a great ecosystem i think what we're maybe not uh doing enough of is um you know shouting about our successes and and putting our putting scotland on the map i mean you have to realize this is an extremely competitive landscape internationally and there are areas helen has spoken about natural language processing robotics that were great at but there was also great innovations coming out of industry you know amazon and canon medical and other kind of leading leading businesses in technology are using ai in their products and i think what we what i personally think might be very useful would be to have some sort of national ai institute that actually pulls all these different strands together and articulates what we what we aim to do in a.i and i think uh you know in the bay center when we talk to lots of companies and investors we sometimes hear scotland is the world's best kept secret uh i'd really like it to become uh the most uh visible and famous place for doing ai um and i think there is another important thing here which is around what does it mean to be a almost kind of a data driven or ai aware nation and i think there we really have to look at how we drive benefit for our citizens and and businesses and government um so there's a bit of a culture to lead from a tech uh perspective and uh kind of it's important to lead from the top and be the best in certain things to get the investment into the country but i think it's important to think about um how this will actually translate to benefit for our economy and our society okay thank you and and helen can a similar question to you just in terms of in terms of the robiterium how well how well joined up are you in academia with you know public sector with business in terms of all those all those elements that need to come together to really make the the changes that you want to achieve through the roboteerium sure um just uh you know uh the scotland is is currently very joined up i was talking to ellen friedman on another panel yesterday and she's from hp in california and even she knew he got the impression that scotland has joined up both in terms of the um academic institutes so we have initiatives like sixer srp who look to kind of join the dots within the academia but also closely linked with um industry with smes um maybe we need to work more towards um you know joining up with the larger players um it's very hard to compete um with these large players in facebook google who have vast amounts of data but i think what the current pandemic situation really has taught us is that we can't really silo these data sources anymore there's huge amount of power to having data all pulled together so that people can work on common problems and common challenges and hopefully that's the way we can move forward and to advance the field and ai and robotics and the national rover term that is as i mentioned that is the ethos is to work really really closely with industry hopefully you know with their data as well um so that we can solve problems um and co-create solutions for them roy what do you think is is scotland is scotland well joined up in terms of that that kind of industry involvement in in sort of building solutions that will help kind of society more generally and what would be your your observations on how scotland does compare to to maybe other countries that cisco's working in yeah david sudden it's a competing world stage yeah something we need to be very consciously unaware of in scotland is that um you know people don't look locally for talent anymore people look globally of where to site locations based upon where they can gain access to high quality good talent um you you see that actually as you look at many of the large corporations where the where they cite their bases um it is a shame we don't have more of some of them actually located in scotland we clearly have you know good talent that does punch well above its weight we are a small country which actually does limit the quantity of that talent that we have which is a challenge that we do have and we need to to face but potentially having you know higher quality talent with smaller teams is a good thing to be able to get our name on that world stage um i i definitely see some fabulous work being done to actually bring people actually to scotland based upon some of the data talent we now have in scotland i would hope that we could replicate some of that great work and data that we've done actually in the field of ai as well to to bring more people there but you know we we must always be conscious that we are competing and we do need to actually continually need to actually punch above our weight to actually you know bring people actually into the country um you if you if you look actually where people have sighted and the quantity of engineers that have been sighted throughout the uk and ireland we probably are underrepresented and that's a crying shame because we have so many good people um so how many highly qualified people actually in scotland that we probably should be able to attract more people and i would i would hope that we'd be attracting a higher percentage um of that investment than us you know has been seen up to now okay steph do you have any comments on that on that talent talent piece you know in terms of when you know when the strategy was put together how obviously all these areas are interlinked ai data cyber security and so on how do you how do you think we do when you were putting the strategy together what were the comments about that that kind of talent base in scotland and how we can develop that as roy's just described well i think it's quite well reported on that there is like a skills shortage in scotland especially around the the data and tech industries and therefore you know obviously ai you know at data lab itself we we have a huge skills program uh you know uh but it's you know it's a drop in the ocean really in terms of what's actually needed as a country um but uh it's it's all part of that you know the way we developed the strategy is we convened working groups uh based on um enabling themes so we did have one on skills and knowledge um but i just kind of wanted to just put in the thing that you know here we are you know universities organizations you know uh commercial organizations talking about ai and driving business and driving investment etc but michael touched on it earlier is that you know as a country we need to also talk about you know the benefits to society and you know everything we're talking about here anyone in the street has no idea about and also doesn't care about and uh and i and i think one of the things we're trying to do in the ai strategy is to try and bridge that gap is that it's not just about our universities being amazing or us having like cisco or amazon or anything in in scotland doing amazing things because the truth is to the rest of scotland no one no one knows you know and and is trying to make sure that there's a kind of whole team scotland approach to this uh so when we uh we put together the enabling themes for the development it wasn't just about skills it was about knowledge as well it's about having a country scotland that is has you know is you know is knowledgeable about ai and data technologies and that there's a huge challenge around that but we need to make sure that's why we ended up having the focus the vision is not about we will have the best universities in the world or we will have the most you know ai companies in the world it's about scotland wants to lead in you know trust where ethical and inclusive ai and that obviously encompasses everything but the societal benefit and how it benefits the people of scotland is a key focus of what we want to do going forward and you know it was touched on right at the start about how do we communicate ai you know it's well said here we all know what ai is and once again like i said you could say to you know what stop someone in the street and then go did you know edinburgh was was one of the first universities that did ai in the world to be like no do i care no i don't and so it's just trying to get that country-wide you know multi-stakeholder buy-in into a.i and and therefore you know raising awareness as a whole and then some of these issues around you know organizations using ai or the skill shortage or people just generally being away aware that there are actually career paths and things in ai is all part of it you know we can't just talk about all these things in isolation it's this massive massive challenge that we need to tackle very interesting point michael do you think do we need to take the majority of people along or do we just need to show them you know show them the solutions and if ai is doing its job under the bonnet do we really need to communicate it or is it coming back to roy's initial point all about those all about those outcomes i think that's a very very um complex question i think we need i mean in the end a technology will only be successful if people trust it and um for people to trust it they must have a level of awareness and they need to be able to make informed choices about how they're going to use it and how it affects them and and how they judge it ultimately and ai has a history of going through lots of cycles of you know when it when it became popular and then it disappeared again and uh we don't want things to happen to ai that happened um to some other technologies and i think our obligation is really to create a culture where the and and this is missing to some extent in the tech industry where really we put the responsible development of technologies and you know we put the the purpose first rather than the ability to do so we're not just going to build things just because we can um so i think on some level you don't need you need the the the average citizen and consumer to be reassured by industry and government and researchers and academics that what we are exposing them to is safe even if they don't understand how it works um internally because that's true of any very advanced technology we can't expect suddenly citizens to be experts and everything um but on the other hand we need to work on the pathway to engage providing pathways to engage with it for those who would like to and and for those who would want to engage with it and i i don't think you know the majority of scotland citizens will become will work in ai but many of them will work with ai so we have to really think about what does it mean you know what is the ai a hairdresser needs what is the ai a lorry driver needs and how will it enter their lives and will they have the skills um to work with it deal with it and feel safe around it and and and and say no when it uh when these things aren't um guaranteed so so i think you need uh and i'm very impressed for example by what finland has done about this i think you need a basic level of awareness that is very broad so that everybody understands what it is as a citizen and then you need the skills pipeline and the kind of the scaffolding between schools universities um business and and the public sector to enable people to access more knowledge when they need it and and and also to um be able to afford that i mean you know it is it is high tech and it is upskilling these things don't come for free so we need to think about also how we invest in that okay um i'm just gonna you took you you touched obviously a little bit on the trust the trust and the ethics there and you mentioned it earlier on steph and there's a question saying in what ways are ethics and trustworthiness embedded in the scottish ai strategy and not just in the strategy but also in that wider sector that is in has got to engage with that strategy and deliver those positive outcomes so where does ethics and trustworthy just really sit in this stuff well i i would say it's the core of what we're wanting to do you know and um you know we've kind of laid out you know our principles and practices and we're following the oecd uh principles um and then we also outlined that we want to adhere to the unicef's guidance on you know as we specifically pointed out the uh guidance of children and uh children the use of ai so you know ethics ethics needs to be baked into it um but you know we're not saying ethics is easy it isn't at all um but the trust issue is really important i just want to go back to what michael said agree with everything he just said before and you know we discussed about it's not about making everyone ai experts you don't have to be ai experts to understand that you know what you're using is safe and we want to get to a culture where you know if people want to become a experts they're easy paths for them to do that that there's you know a pathway for them to do that or people just want to find out what ai is there's a path to do that we're not going to force educate everyone you must know exactly how machine learning algorithms work but you know it's kind of happened like um you know we reference finland's elements of ai mooc you know on our website you know it's there it is it's you know it's slightly marketing gimmicky but it's brilliant you know because you can get uh it's a free course for everybody and we just want to make sure that if anyone wants to find out more there's more but we want to create this culture that you know if something you know if some if an ai technology is developed in scotland and as you uh and is available for use by companies in scotland that there is an element of trust that this has plays you know adheres to a certain uh you know um range of principles uh that you know that this is not a dodgy thing you know and uh i articulated that really badly but but uh that's like that's what the ai playbook hopefully will do what what's your perspective roy from you know from it from a business perspective as to how how embedded ethics and trust are in in in how ai is is developing across a whole range of partners yeah david it's a you know it's an interesting question because um you know we talk about ai actually as as as a topic and yet ai actually covers a number of categories in fact so if you look at you know just you know basic you know data science data analysis um you move forward from that into just you know machine learning um you know potentially when you look at your basic data analysis and machine learning that there's nothing really challenging around the ethics around those they are purely algorithms that are run on data that come out with uh you know a predefined actually output there isn't really effectively then the challenge that people potentially have is around the intelligence aspect of actually making more independent decisions actually they feel actually with artificial intelligence so you know from my perspective actually and you look at the field of artificial intelligence there's quite a lot of actually technology that sits within that categories that are you know relatively ethical and very simple to implement actually with off-the-shelf tooling that's very very well understood that can provide you know massive benefits to you know people and organizations i think it's really important to move on to some more of the you know the more advanced types of our official intelligence um algorithms and um potentially in other forms of use cases and that people need really more um support around their understanding of what that can be you know if you take um you know there's some there's some you know really some very you know leading edge fabulous utilizations of artificial intelligence you can use nowadays um you know that you know to me actually one of you know the fabulous examples really actually sit actually in health when you look at actually things like you know cancer analysis and you now have um artificial intelligent algorithms that can actually detect things actually in a medical environment to as good as the best people on their best days and i think those you know things those things actually um are you know fabulous neighbor enablers to technology particularly when we have a lack of medical staff now actually that just simply just simply cannot cope with the quantity of people that need to be looked at um but there's those types of areas where i believe you know ethics and really understanding what is being used are areas that we needed to provide a lot of support to people with um you know if i'm going to open up an app on my mobile phone i'm going to work out whether it's quicker to get into the office actually by hopping on that bus or getting that train today or not i probably am not looking for a ton of really support actually around that now am i and you know that's where you know machine learning actually just works fabulous like you know we can simply look at patterns um across a period in time we can look forward into where other people are paler pulling some data that you just can't see and say you know what that bus is going to be faster for you to get into the center of edinburgh today hop on that one so i think it really does actually know vary based upon you know where where it is you're using in the technology and how much level support you need to actually provide to people to give them an understanding of what it is that's been utilized yeah and there's a question in the box just going back to the data that you use all ai requires constant flows of data which raises questions about privacy it's not the ethics is not just about the actionable insights it's about where the information comes from in the first place michael any any comments on that please so i think i think the the data issue is very important here um but i think it's important to also recognize that um with a lot of the problems we're seeing around privacy and around the information sensitive information leaking or being misused or resold and so on i think the underlying issue is very often kind of poor uh business practice and uh and you know although the incentive for doing bad things with data might be because you're using ai it is not kind of the ai's fault i mean i think we we uh need to always remind ourselves that um while ai um raises a lot of kind of imagery around some magical thing that does very clever things um almost uh in every case we've seen it's really kind of still uh performing very narrow specific functions which are um very carefully crafted by human engineers and and whether and where the people in organizations who decide to deploy these applications really make very can make very conscious decisions so so i think um we need to think about um these things to some extent separately what is an issue is that a lot there is a lot of kind of opportunistic data harvesting in all kinds of domains um which is driven by you know a very a very kind of opportunistic way of thinking of well you know we could do something very exciting with it so the more data the better i think i think we do need to get away from that narrative and we really need to think about the value of data in the val expressed through the value of the applications that that are built on top of it um some of the examples roy used uh for example um you know you can do a lot of the stuff what i call ai in your pocket a lot of the stuff that's on your mobile phone you you can do a lot of that analysis without knowing anything about the identities of people for example um and uh and i do think we need a bit of a change of culture around that um and and really the public to expect that data should be used to the extent that it's really needed i think there's also a bit of a problem around the new technologies we're seeing in machine learning and neural networks and so on that they're increasingly data-hungry when they're trying to solve more complex problems so i think there is really also the onus is on us as as researchers in academia and in industry to also look at new methods that can do certain things with less data okay thank you um helen i'm just going to come to you if i can ask the panelists try and answer these questions quite quickly now it's only got 15 minutes left so now helen if you're recruiting somebody for harriet what does it matter that they're in edinburgh or that they can get to you or would you recruit from anywhere now would you take someone from anywhere in the world because you can they can just do the job remotely so that's a interesting question so as you've seen in the pandemic we're all working remotely um but for example robotics it can be quite tricky um but now we're seeing the use more of digital twins which is basically a data representation but it can be a digital twin of infrastructure of a building um of you know offsite of a factory so there are new innovations coming out because we're all working remotely and that we can that we can really leverage um i mean i would always prefer to work side by side next to somebody especially if i'm recruiting somebody new so there are there are challenges for recruitment and visas etc but the world is definitely definitely opening up roy just briefly on that point would you do do people need to get together in the same place or can you can you work completely effectively remotely is is the world a talent pool now yeah so that's a great question it depends how how how widely open you would like your organization to be you know we we think you know moving forward there will be a hybrid works workspace actually for that and um let me just just give you a great example i was a couple of minutes actually late joining the panel today actually because then i was actually um collaborating um on video and white boarding was a colleague in the middle east working on a project for deliver sun for a customer in no man and i'm based in scotland and the rest of the team that looks after the customer actually are based actually between manchester and london so you know you come back actually and ask me the question do i need to be there no do i like to be there sometimes yes and we think that you know there will be a hybrid that balances between these environments where you know people will need to be in an office um we like you know we are social animals we like to get together with other people it helps when you're collaborating um to be together with people but do we need to do it all the time like we used to perhaps not are we always going to be 100 remotely perhaps not but um you know for for people that need to drop their kids off at school early maybe they'd like to work from home two or three days a week where people need to collaborate for instance in advertising agencies maybe they'd want to be in the office to get all the teams together to collaborate in the same room so i think it will actually change going forward i think the only thing we can actually say based upon actually what we've just come to recently is things will change and i don't think any of us actually have the true answer of what the outcome will look like in a few years but i can definitely say it it won't look like it what it did a couple of years previously okay thank you roy um michael there's a there's quite a complex question here about engaging ai and and robotics as well with the arts the culture sector and literature arts departments and universities and looking at this called trust public engagement issue how you know how can you elicit you know that sort of the artistic community and writers and so on to to to get across that message i don't think you see the question in the box it's quite a complex question if you can have a quick go at that michael you've got 30 seconds to answer that question i'll try so we have a actually a big program in at our university that's called creative informatics and that looks at kind of you know help using data and ai to innovate in the creative sector and also there's of course lots of work with digital humanities people in literature and so on trying to use data trying to use ai for their work um so i think there is a this is a very fruitful um interaction in particular because the arts and creative expression allows people to experience and engage with new technologies in ways that are um novel and that may maybe communicate a lot of things um to the spectator uh that uh we can't do just with words or we all might be too complex if we try to do them with with lots of um kind of you know scientific lingo um what i would be a bit careful about we've seen amazing artworks being done with ai for example is a great great example of uh an artist using neural networks to play around with the um gender perception using uh imagery from uh uh creating artificial kind of uh drag queens um which is which is extremely uh fascinating i i think what what i would um caution a little bit against is that while the science fiction element of ai is extremely inspiring and stimulating for debate and and and really it's it's really about human nature and the future and humanity and the future of humanity i think it's important to um be very aware that that's not the reality ai can't even you know answer my email or load the dishwasher yet so um you know it's always great to dream up future visions um but i think the the focus now needs to be on making best use of what we have and developing that further okay um we'll just we'll just look forward you've mentioned their novel sort of exciting application i'm just going to come to all the panelists in turn staff you first just so you're aware um what what to you is the the the big exciting thing what excites you most about the potential of artificial intelligence what what's the big novel application that you think oh that's brilliant that you know if we really nail this ai can do this amazing thing you know what what is it to you steph first of all then i'll come to helen roy and then back to michael uh for me uh with my other data laptop uh hat on is uh is the potential in healthcare um ai has huge potential to transform healthcare um i'm i'm not saying that you know about uh you know ai taking over the jobs of radiologists or anything it's about ai augmenting and insisting you know an undistrained health system to make things better and quicker and better for patients uh there's so much potential there um you know in medical imaging is the obvious thing um but but yeah but i'm a big pro of human centric ai that you never remove the human from the from the equation there so uh so yeah that ai technology works with medical staff to provide better patient care uh and that's the most exciting thing for me because i think uh that that's a real game changer there helen i'll broaden it out to kind of ai and robotics for you you know what what really what really excites you well just going on to uh carrying on from the human-centric angle that steph mentioned so conversational ai already making massive um leeway on that um so if robots and ai algorithms could communicate as well as humans they could explain their reasoning if needed if so the user could probe and ask them why why are you making that decision what you know what's the reasoning behind that i think that will be key for adoption for both and ai and robotics going forward what what sort of areas helen i mean we've talked about this in in previous discussions you know you've talked about health care as well in terms of ai and robotics and also energy the energy sector and so on can you just give us a couple of for instances where you can already see ai and robotics making a real practical difference and delivering much better outcomes sure well um just take energy for example if we're looking at robots that can do continuous inspection and maintenance but it goes back to robots being able to adapt to dynamic environments and in order to do that they need ai they need vision components they need to be able to sense they need to do obstacle avoidance and they need to be able to communicate their findings in a way that's easy to understand and so that's one example of how ai is moving forward in the energy energy sector and about trustworthy autonomous systems there's a large uk wide program um coming from ukri it's about 34 million to look at trustworthy autonomous systems and that includes ai as well and in scotland we have two of these nodes so these smaller projects parts and the wider program and looking at trust um human interaction for trust and also governance and regulation led by embry university so trustworthy ai trust with the autonomous systems um it's absolutely key going forward and i think the government is realizing that as well okay roy what really what really excites you about the future what gets you up in the morning and thing are you you're doing something on the day and you think this this is this is brilliant this is gonna make the real difference you know ai helps us in so many areas going forward but you know fundamentally it's all about data and that's the one thing that's been you know happening actually over the last very number of years is we've just generated more and more and more data and and you know we have the potential now to collect more and more data and the challenge that we see with that is that people cannot then understand it and we need something to look at that data for us you know we we spoke right at the very start i said you know what's the applicability of of of ai and i said it's really about the use cases of using ai and there's been questions about of you know the ethically of the data itself but you know just imagine you know any of us who may want to go to the sun again in the future think of that a beach and a cold beer or someone bringing it to you think of that and you get there on a plane what those planes got they've got these massive jet engines how do we know that those jet engines are safe and secure what you do is you sense them up you put hundreds and hundreds of thousands of devices and sensors on those jet engines and you turn them up and what does that do that generates so much data um using ai we can then process that and say that engine is a safe engine i can sit in the beach with my pina colada in the sun and that's you know that's the type of thing that you you want to be able to see we want to be able to process the vast quantity of data we all generate and actually gain insight and benefit out of that and to me that's one of the things that ai is going to give us going forward in the future okay and michael what what you know what really excites you about about your job you must be doing so many different things on a day-to-day week-to-week basis but you know what what's what potential really really excites you i think i mean because because i have seen what people do and you know we we've got kind of 50 companies around us who do lots of different things and we work with lots of other researchers around the world and so you've seen we've seen all these great ideas they have in things like health and manufacturing and retail and so on and so i really like to think about the areas that people maybe haven't thought about that much and i think one i think you know on a global level the next big challenge is going to be food um we know that that i mean in conjunction with climate obviously um and i think there are actually things we can adapt from from other areas like health which we could where we could we could do great things there you know i'm thinking of of the farmer in africa who could take a picture of his um animal and and find out what their um uh what disease they might be suffering from or what their nutrition status is and you know we've got a long history of doing animal genetics here and you know based on those models you could very easily help increase production but there are other areas we can understand uh disease and crops we can try to reduce the energy consumption and carbon footprint of food production we can improve um supply chains and uh i think um you know some uh people are already getting into these things so we've got a a company that monitors uh uses robotic devices to monitor grains in silos um but there will be other things like uh understanding we've been talking to a big uh global coffee chain you know how does the how does the bean get from africa to the customer somewhere in the us or in china and you know how we can how can we actually optimize that to um environmental and economic um outcomes so uh i'll have to stop you there michael because we are running out of time i presume you were talking about actual silos rather than figurative silos there which we do talk about a lot in this area final question to everybody step first single action 30 seconds each single action to help us join up data driven innovation ai more effectively to deliver positive outcomes oh um first uh well i i from a you know ai strategy uh but please get involved you know get in touch get involved join events and things when we put out consultations events get involved just you know talk to us if we haven't approached you to talk to you please come talk to us uh because we want to make sure this is collaborative going forward uh you know no no small group can solve all the challenges for everyone we must all work together okay helen let's store my words i was going to say collaborate um so all of these projects going forward need multi-disciplinary teams and you know no matter how big these institutes are i think cross-institute collaboration is is going to be key going forward and we already have a lot of projects um that do that so um hopefully go forward let's continue to collaborate roy yeah david you know what i mean we think the future you know what should people be looking to do so like we think people will be you know reimagining all their applications how they actually use their applications we think they need to secure the data across those applications through all of it we think they need to empower their teams to move forward to work differently in a new hybrid way and i think you know when we start doing all of the above we'll start actually seeing the other real benefits out of the underlying technology to all those different organizations that take advantage of those transformers technologies in the uk okay and michael finally you know one one action we could take to join data driven innovation now more effectively i would see i would say work closely with the decision makers in organizations to help them understand how they can put these technologies to their service i think that's one of the biggest roadblocks um to wider adoption and and collaboration okay thank you very much there was one fiendishly complex question in the q and a box which i'm going to ask any of our panelists if you can answer it afterwards that would be brilliant a guy who's been juggling his sick baby he adds not literally juggling his sick baby in brackets which is helpful he's asking a very complex question about defense if anybody thinks they can take that on or pick it up with him offline that would be absolutely brilliant so thank you very much to roy to helen to steph and uh to michael uh we've covered we've covered a lot of ground there it is a big complex topic but i think again it all comes back to everybody joining the dots working together and collaborating effectively and actually looking at what is coming out at the end what problems are we trying to solve um and how can we work together to get there to best effect so thank you for listening in uh look at the ddi website there's lots of extremely interesting stuff going on and there is some great collaboration thanks again for tuning in thanks to our panelists and please enjoy the rest of your day cheers
2021-05-28