AI UK leaders reflect on findings of recent RAI UK consultations across the devolved nations and contrast these findings with the AI summit outputs.
Featuring:
Martha Lane Fox – Chancellor of the Open University
Prof. Chris Johnson – Pro Vice Chancellor for Engineering and Physical Sciences, Queen’s University, Belfast
Prof. Yr Athro Matt Jones – Director, Morgan Advanced Studies Institute, University of Swansea
Professor Sana Khareghani – Policy Lead, Responsible AI UK
Prof. Sarvapali (Gopal) Ramchurn – CEO, Responsible AI UK
We’ve got Matt we’ve got Chris Santa and goo great hi guys well it’s great to be here and I am so thrilled to be able to talk about um the word responsible in this I know a little bit about responsibility technology and I know not anything at all about AI so you’ll be
Relieved to hear I will be like this and just hearing from all the exciting work that has gone on with our regions uh in in the UK this Summit is clearly where the action is what I’ve been so impressed by today is the focus on outcomes and things that feel achievable
As a kind of practical person and someone that likes to say okay now what it’s fantastic to hear what is happening beyond the kind of existential Frontier world that um is also going on in parallel to this um I’m going to hand straight over to gopal who is going to
Tell us a bit more about UK R AI um I think I got the ini right and all the amazing work that’s been going on thanks thanks Mara and um great to be here on this panel and to coane this amazing event uh thanks to to the to the whole
Team the French team and also very strange to be on a panel with you Mar because um I remember back when I was um an undergrad at Imperial years this is making me feel 870 years old the first startup I invested in was lost.com so oh [Laughter]
Sorry um it’s not that bad um so responsible auk so I’m the CEO of responsible auk and and just to talk talk about the name responsible AI we we initially thought we’d call it AI UK because we think all AI has to be responsible why should there be a
Difference between responsible Ai and and and other AI I think that’s what our mission is is try and make all AI responsible and that being part of the conversation so responsible auk is a 31 million pound program funded by ukri UK research and Innovation for those who
Have you who don’t know um it is not the first investment in responsible AI UK has invested over maybe 200 million pounds in in Ai and really responsible AI Technologies over the last uh 10 20 years um I I still run the trustworthy autonomous systems Hub which is at the
Center of a 33 million pound program that started three years ago uh but responsible auk was funded earlier this year started back in May and uh brings together an amazing leadership team including people on this panel um s Matt and Professor de Wendy Hall Professor de
Muffy CA Tom Roden um proas scupa from Kings College London um May O’Neil from K University Belfast and what we try to do with this um leadership team we call it is trying and bring those different voices from across the UK um to make sure that we engage in these National
And international conversations around responsible what it means to to to the designers to the developers uh to the the users of AI um funding research to address the key challenges that all these stakeholders all these all these parts of these AI life cycle um face um and delivering that research to Industry
The output of This research to Industry to policy makers and also try to translate those into skills related activities so the has this this program has ambition to do to do all this research and and to deliver these outputs over the next five years so started back in May we’ve already um uh
Funded a number of projects which I’m going to talk about at the end of this of this of this session um but uh we are there trying to bring together the whole ecosystem trying to glue that ecosystem which we think has been a bit um uh split up because of the different maybe
Emotional words used around AI safety around responsibility around um uh regulation can we bring all of these actors together to talk about responsibility to do investigate the key issues of responsible Ai and address them uh meaningfully fantastic thank you we’re going to hear a little bit from
All of the regions I’m anxious by picking one person ahead of the other then I might call some kind of destabilizing to our United Kingdom hopefully not um and then we’re going to have some time for questions from the audience and I think at the end of the
Day uh otene Li is going to come and tell us about ukri so that’s very exciting end point to the summit so s do you want to start well you picked on me I won’t have any region to talk about but I I represent
London of London yes um but no for my my um my kind of the the reason I’m here was to talk a little bit about um how policy makers use the uh the research on Innovation that is created by our universities across the regions to help them make better informed policy
Decisions which I’m happy to come back to but we can move on to the regions and okay back well why don’t we start with Northern Island I’m GNA pick one let’s go with Northern Island okay so first time I’ve been referred to as a region I am Scottish but I’ve lived in
Northern Ireland for the last three years um I uh I come from a different area and I and safety to me is something totally different to what it means to most of this audience like a I’m an expert witness in the communications failures at grall and prior to that I
Worked on human space flight um I and right now I work on machine learning and weapon systems and I I believe that some of the things that we talk about today in particular ethics should actually be changing the way that we do engineering so I think engineering
Should be based around the concept of ethics and that Safety and Security are subcomponents of that and I’ll come back to the person that had a little go at Great Britain I think that if we have that confidence that ethics provides that we that we’re doing things for the
Right reasons then maybe we should emerge a bit more in terms of global Britain for the future and I and I’m a firm proponent for those of you that are policy makers of the integrated review and I really do think we need to have ethics to help us regain confidence as a
Country going forward and do you want to tell us a little bit about what’s been happening in Northern Ireland um some of the work I mean not generally in relation to this we could be we could be here for a long time being liveed the big questions yep
Uh I mean northern Irish industry is largely based around small to mediumsized Enterprises and there’s an awful lot of fantastically Innovative and creative firms there um and I I think the thing that the thing that holds us back in the space a lot in the UK is the absence of regulation in
Safety and Security so in most of the industries I deal with you’re not allowed to put machine learning into those applications because right now there’s no mechanism or framework for regulatory approval so just put very simply if you put a machine learning algorithm them into a safety critical
Environment how can you be sure that it won’t learn an unsafe Behavior so there’s phrases that you’ll see coming out on Wednesday and Thursday which I disagree with so people will talk about safe by Design and secure by Design I have no idea how you do that when the
Threats change and when systems by definition learn so I think one of the things I would come back to the philosophers and some of the others that have spoken here is that we need really hardcore engineering to point out to people what is and what is not possible
Today thanks thank thank you Matt tell us about what’s happening in Wales what’s happening in Wales so um first of all I just want to address what responsible AI is we’ve been here all day uh when I was growing up my parents told me you need to become more
Responsible and to me that sounded really boring and really really constraining so I rebelled against that so the key thing to say is responsible AI is not my parents definition uh what we learned in card if we at a town hall I think a couple of weeks ago we had
About 80 or 90 people from across public private third sector government coming along to be part of that National conversation that gopal has convened in the responsible AI UK program and I think what we landed on was if you take responsibility and just look at that first bit of the word response or
Respond uh for us in Cardiff U maybe because we are a nation or principality or an oppressed people depending on where you sit um music and song is very important to us so how does this new instrument of AI respond and resonate with our values and I think there was a
Really lovely phrase that one of our panelists earlier today said how will that let the souls of our children sing and that was a lot of discussion um in Cardiff it was tempered a bit Martha with you know way has quite a good uh long history of dealing with materials
That have gone out from that place and forged in revolutions you know coal iron where I come on Swansea copper but of course we also have a very long uh experience of that being an extractive an exploitive a very dark and a very dangerous time so the discussion in
Cardiff was okay how can Wales with that long history with the desire to let Soul Singh play a role in the future which is AI and I think you know what we have in Wales is we’re devolved power we had a members of sen of at the Town Hall and
We have things like the future Generations act if you haven’t come across it look it up this is not a party political point but it’s a bit of legislation which requires all of public policy to consider what what will be the impact on your children or your great
Grandchildren we also have a social compact legislation which requires workers to be involved in decisions about the future of Technologies and we heard earlier today I think you know resonance is there that if we’re going to do this right then it needs to have every voice as part of the conversation
So during that day in Cardiff there were lots of different perspectives but a real heart to ensure that people people are placed at the center of this endeavor so it was a very positive day lots of fun bit of singing uh and quite good food actually oh excellent we
Haven’t talked about the nibbles Regional nibbles um are you going to address yeah so uh we ran a number of town halls as was mentioned so Glasgow um Cardiff uh London and um Belfast sorry Bast Belfast yes um and we could see different concerns being raised or opportunities it’s not always about
Concerns but new opportunities for example in Northern Ireland as as as Chris mentioned this huge commune of smmes um and and this interestes in regulation and regulation technology being one of the key areas where you have a number of companies that need to work across different jurisdictions very
Closely across the border so can they be a center for this uh regulatory uh Technology Innovation um in uh Glasgow and London I think lots of uh concerns were raised around labor rights around around the future of work um and uh where there are questions about whether
We should when people lose their jobs should we turn to Universe basic basic income basically to to to to fulfill that that that to fill that Gap um and as part of that some of the members ofina NE for example is part of this the TU task force looking at the future of
Work are working closely with the institute for future work as well so lots of new relationships being built up as part of these Town Halls um we’ve uh also got lots of interest in in in sustainability um two Town Halls we we had these discussions around AI for sustainability and sustainable AI
Because AI now is starting to feed a lot on data and data centers consume lots of energy um is it a good use of our energy uh sources and and we’re starting to reflect on these themes and trying to shape our funding programs to address some of these
Concerns I was really struck um I’m president of the British Chambers of Commerce we have 100,000 members and businesses and I’m lucky enough I get to travel about and do fun things like go to Sugar factories or go into um Advanced manufacturing I even went to
The biggest port in the UK fix recently all my Logistics uh dreams coming true as everyone thinks in my family I put everything in boxes but the thing that has struck me is just to your point about the kind of um disparities between both in regions but also in sectors I
Was in Doncaster and there was a round table for businesses there and the just just so happened to be two insurance companies around the table not huge businesses but fairly chunky normal smmes and one of the guys who was running and they were both guys one of
The guys said I’m planning on replacing half my team with better language models in the next two years I imagine I’ll be half the size by 2026 and I was like wow didn’t expect to hear that in a back room in Doncaster and then the other
Insurance guy looked at me and you know this is not critical of him and said what’s chat GPT and it really rammed it home to me how not only are we creating two track economies in different bits of this country but even in these microclimates we’ve got these very very
Different things happening within sectors yeah yeah maybe you want to come in yeah I mean I I I wanted to just add one thing that struck me and I was at three of the four Town Halls was the amount of work that’s being done in each
Of the regions as well so the the the uh the the research The Innovation the types of challenges and problems that are being addressed um I just thought it was it was fantastic ability to kind of see that richness that’s happening that’s being very targeted to the specific challenges and problems that um
Are being faced within those um those areas but yeah I mean to to to the question around disparities um it was one of the one one I had a meeting when I was um when I was heading up the office for AI with um some people from
Some leaders of kind of uh grocery stores um and large retailers say and uh you know they sat and they were also all men um and said oh we use AI all the time you know this is this is not new for us we use it we we have these little cards
It tracks everything we move things around based on it you know and that was great and so I asked them um what their what their um waste was the food waste and they said well actually you know ours is better than the national average is around 21% or whatever um and I said
Well okado um which is an AI first company their food waste is at 1.2% and it’s actually apparently it’s even less now um and and and that was a big kind of moment of oh you know because you don’t have to actually even go any further than that because it’s very
Obvious what that equates to for them but I think there is there’s a real question here when we say that AI is being adopted and I hear it all the time about how people really worried they’re missing the boat on this that actually adoption of AI isn’t anywhere yet right
And um and terms of the the level of adoption um and and how much whether it’s a Bolton versus you you’ve adopted it you’re you know considered an AI first company so it’s it’s running your entire business line end to end is very very different can I pick up that um
Disparity I’m Going To Go’s point about this being an international um collaboration um quite a lot of the work I do is in the in the global so when I was in South Africa about a month ago and two things struck me so here’s the negative side and you can look this up
Yourself uh it’s horrific it’s and it’s also very obvious uh but if you go and that’s I did into townships in Langer and Del and kyisha and did some workshops and asked people to put in prompts about um into Dari the image generation or chat GPT and to see what
Would come out um so people typed in things like Show me Life in Langer and it showed um happy smiling white people enjoying themselves now why is that well it’s because most of the photos are coming from people who do what’s called and it’s a horrible phrase I’m sorry
Slum tourism however on the much more positive side the people I was working with there could see how they might use some of those Technologies to create new businesses themselves and even more insightful gave us challenges which undermine some of the fundamentals of foundational AI so for example CA it’s
Spoken by oh millions and millions of people it’s a fascinating language I guess some of some people in the audience know it um it’s underrepresented in uh artificial intelligence systems but because of its structure it ask some really serious questions about how you transcribe to train the model and how you accommodate
The kind of code switching changes and that can’t be addressed just as oh we just we’ll just put a better interface on requires anyone who’s doing research to really think about how you build and test those models so I’m absolutely for international collaboration not simply to make sure everyone’s involved but
Because they will disrupt some of the fundamental models that we will use um in the future so that’s a key part I think of what raai is going to be doing yeah it’s a super important Point can I ask a slightly challenging question to anybody that wants to take it do you
Think that there will be a international or even UK um collaboration around what responsible means and is that important or not I think I think there many many institutes organizations being set up we’ve heard of the AI safety Institute we’ve heard of the UN AI advisory um Council we have general partnership on
AI a number of of these bodies are coming together to try and Define what responsible is um for me it we didn’t try to impact the term responsib because it means different things to different people it was specifically chosen to avoid people bucketing it into a technical or social science sort of
Research realm so um it it needs to be it needs to be a broad conversation with everyone and what we are going to focus on is is where are the research questions where are the big nuty research questions that we need a multidisiplinary perspective to to
Address those um but and ready to work with all these organizations to try and answer the questions I suppose my job is to try and give a bit of a reality check here um I think within the political context we can’t assume that every country around the world shares our view
Of what responsibility is and I I think I mean I I’m also serve on the national cyber Advisory Board and I’m struck by um things like plausible deniability of involvement in cyber incidents undermines protestations from certain governments around the world that they don’t actively engage in in very offensive cyber activities and and
Um I I was asked to give advice to the Irish government on their review of overseas policy and and qu the question they were asking was do you believe that um we should view developments in machine learning and cyber as something that is possible to regulate internationally and I said the only same
Way to behave is if it isn’t because if if you act within a constraint set of parameters what you’re asking is 18 19 year old men and women to go and face systems that they have no no no means of of addressing or combating so we at least need to be
Aware of the possibility that there’ll never be agreement or if there is agreement that the reality behind how nation states behave is not in accordance with how they behave here equally I think we should hold firm to our ethical standards and the our ethical standards should include things like Effectiveness which which means
That will’ll then have the confidence to be able to deploy things knowing what the effect of them would be and and being confident that we’re making the right decisions and that’s a li that’s a within machine learning that’s a life cycle approach and it changes the whole
Way that you view engineering so for example I think into the future we will have ethics Duty holders for a product that starts from the grave until retirement and includes sustainability security and safety within the duty of of of that of that corporate obligation throughout the life of the product that’s very
Interesting do you see that happening anywhere why do you say that with such certainty because I think it’s happening in the military and they’re starting to really ask these questions properly because they’re right they’re right at the the crusp of this but it’s not not talking about existential Killer Robots
I’m talking about people have to prepare as if that’s what other people are doing yeah and I think the only so already for instance within mod you tend to have a safety a safety Duty holder and I think safety is one small component of this larger concept to do with ethics which has
Fairness transparency Etc within it but those things that I say I can see in the transition in the military are equally true for industry so you you could talk about a finance company and you can see it within the you know kind of ethic sustainability and governance yeah but I
Think personal responsibility for those is really important if you if you you know if if you’re if you’re not clear about one person at the end of the day having having a sense of direction over that within the company then it’s meaningless can I just also come on in
This um I think one of the challenges with what what we see with AI applications is that the connections to very complex Supply chains of AI yes AI being trained on data sets that are provided by universities or by people who’ve collected them just put them online um webites used by GPT for
Example to train itself um which are being curated sort of uh filtered by some people so there are other people involved in the filtering process in the in the in this in that whole supply chain and increasingly we’re going to see tools that they’re building on these
Complex Supply chains so when you talk about you know ethical AI how far do you go back down the supply chain to check on this yeah right do you go back to to Kenya where someone’s classified an image as you know being uh OB seen or or
Or or or or okay um who do you go back and check um with when it comes to valid ating the ai’s performance the accuracy Etc across the I think that’s the big big challenge we have with this highly interconnected real time Supply chains it’s very different from from many other
Industries yeah it’s really important s do you want to tell us a bit about how you’ve seeing this link up into the kind of policy ladder if you like so I mean I think what a couple of these things we we’ve spoken about or has been spoken
About today and I think for me um the most important bit of getting policy right and policy is um is complicated and and you know we have to think very kind of I mean to overuse the word responsible again but you have to be very responsible in your approach to
Policy because it has a lot of kind of effects on on the entire populace um to try and to try and approach it well we need to have the perspectives of lots of different groups um and when we get it wrong is when we listen only to the
Loudest or um the the most well-funded or you know the ones who are paid to come and talk to government or whatever but um you know and that’s where having a group like the responsible AI UK program allows uh a a representative group from another perspective to offer an opinion
Um and to to to do the research and to be able to to kind of come back with backed empirical evidence in terms of here’s here’s this other perspective now that isn’t to say that the voices of Academia have to should outweigh other voices right but there should be an even
Kind of group of voices that policy makers listen to one of which should be Academia one of which should be industry one of which should be civil society um and so on and that that then allows us to kind of make a a nice well-informed policy decision um you know every one of
Those groups have have their own uh their own incentives for for saying what they do or their own incentives for for for pursuing a certain type of research so we can’t hold one above others but it’s important to have that kind of varied voice and I think what
Responsible auk does and is trying to do is to be the convenor of that voice from an academic perspective um and also to so more so now than ever before make sure that the voice of or at least that the R&D that’s happening in industry is also included because um
What we’re seeing these days is there is an imbalance in terms of where the money is and where the labs are actually doing the research right and they they do get a a quite a bad rep um and sometimes you know welld deserved sometimes not but the the research is happening in in
These companies now um a lot of the research and it’s important for uh there to be a a conversation to understand what that research is what what it’s finding to to kind of poke and prod at it from a different perspective and and so from um from our perspective
And I’d love to hear the panelist perspective as well um is that you know when we talk about responsibility we want to include that voice as well we want to have that voice of industry and the the the best organizations that are doing this research to be included when
We when we give our advice I was just going to say one of the things that concerns me a little bit amongst all of the things that I’ve heard over the last couple of weeks in preparation for the Summit is the lack of uptake of machine learning in government and there so many
Problems that we’re facing today that would be amable with better data better curated uh and better data that we have greater confidence in from multiple data sources and and I think one of the things that I’m coming back to the focus of this panel session about responsible
AI research is it’s not just about responsibility in the artificial intelligence itself I think the responsible view is promoting the use of AI to solve society’s problems and I don’t think that’s happening a very specific example is I don’t know what the other panel members think but I have
Zero confidence that we’re going to meet our emissions targets and if that’s true then we need to prepare for the consequences of extreme weather events and the coroller of that is that you can already see patterns in the UK where fatal accidents are occurring because government and Industry are failing to
Prepare for the for what you see around you today and you know storm Kieran coming in you if you look at the uh abedine Stone Haven accident that um the infrastructure company admitted liability criminal liability for last couple weeks um I think we’ve got difficult times ahead and we really need
To leverage UK infrastructure to be resilient to to those changes and the focus on the impact of AI on emissions and so on is really important but we also need to strengthen Society so that our children um are able to resist what what might lie ahead and I don’t think it’s I
Think it’s reached the point where we can be pretty confident the effects are going to be fairly longterm and fairly extreme yes well want to pick up on pick up on the um the children I got three children they they’re old but they’re still my children as I keep telling them
Um and I certainly had to pay for dinner for two of them last night so they’re definitely my children um there is a lot of if you talk to your children if you have them or or or younger people there is a lot of concern out there they’ve
Had to deal with an awful lot haven’t they and there some of you in the audience are much younger than me the last n years has been horrific and now you’re hearing a lot about the the threats not the existential threats um but some of these ever threats that are
Near and present I think you know one one thing a research challenge i’ put out there to well certainly to myself and others on the on the panel and to yourselves is to think about how you can give younger people tools that really allow to them to channel not just their
Intelligence I think there’s a big problem with this focus on intelligence go back to the musical instrument analogy I started with think about your favorite um violin player or I don’t know harpist or whatever it is they are channeling yes cognitive physical emotional spiritual energies through that instrument and they are making a
Difference so the research challenge is the positive thing the most responsible thing I think we can do as AI researchers or anyone involv is to really think about how do we go beyond this notion of an intelligence and allow our children to amplify their god-given talents and then that light will
Outshine all of the darkness that we see around us for a big problem and it could happen this week that people get you can’t sing your way through climate change right can I but if what is the point of saving the planet if there isn’t any Joy
So I was just thinking please God please God let it amplify the talents of my seven-year-old trombone two two points to up this one can I can I just jump in so in terms of compute and the energy use of of of uh of data centers I’m not
Sure you know about this but data centers now account for more energy assumption than the construction and Maritime sector Y and Airlines and Airlines and part the road and Who does these data centers serve right do they serve your Facebook feed do they serve your Netflix streaming application or do
They serve the needs of the most needy in society right so we need to come know we need to have a policy around you know what is the worth of a computer cycle to society or or the worth of a of of a bite of data to society because that’s
Costing us energy and most of these companies actually take the energy from from renewable energy sources that’s what they claim at least now they 100% green apple claims this Google claims this but when you say these companies mean the data centers store the data that keeps our Energy Systems going it
Keeps the lights on that keeps the NHS going and government is one of the largest procurers of data services in the UK so it’s our data it’s us it’s everything in society it’s not Google versus everybody else it’s it’s the fabric of what we rely on sure but there
Are some applications that do drain more energy than than those government services and that’s why I’m thinking we we need to have a a saying in how these this um how we choose to tax maybe these applications based on not on the carbon footprint that they have but on the
Energy to draining away from more meaningful uses of energy um that’s one one point I wanted to raise the other one was about um uh ensuring the future uh future Generations are happy or knowledgeable and and and and have an enriching experience with with AI so I
Like to bring up the example of my two children um who are learning to play piano using an apple that’s AI powered they are learning uh languages from du lingo which is also AI powered and uh me too I’ve been learning French all year
And I have to tell you I’ve got worse so I was going to come to that none of the people I know who use Jingo actually speak the language that day I’m really good at answering questions and getting higher up the leag exactly so the way
The way these apps work is the guy who invented it was vanan he he build a capture he’s he’s really good he’s really good at gamifying experiences and it’s all about playing a game it’s not really about learning the language in the end I mean my kids game the system
All the time just to get a few XPS to go up the yes you know you know the story with simply Pi is the same thing um simp piano sometimes well the way it works is that it it it plays out the notes of whatever song you want to to play um and
Doesn’t it doesn’t register U multiple key presses that’s I I found that out when I when I was watching them play but it it get better and he learned the language and everything but what they are again getting used to is the idea of gamification of rewards and is that the
Right thing for for them right do do we want people to learn something because okay well we’ve got lots of these are good big questions we’ve covered a lot of ground from singing in Wales to the future of the military Warfare and now to gamifying in my French lessons so can
We open it up to the audience we’ve got a good six seven minutes for questions and it would be great to hear from all of you if you could say what’s your name and where did you come from and I will be S the black here on the stage um yes
Please gentleman here in the middle you’ve whipped your hand up very quickly thank you um thank you very much for a very interesting talk um my name’s kush kodia I’m a social entrepreneur and I just want to link the talk of what we’re having right now with
The current reality we’re facing in the world um so one of the titles I wanted to highlight is never in oxfam’s history have we seen a humanitarian crisis like the one in Gaza and this is from the CEO of Oxfam so why should the rest of the
World trust the UK to be a custodian for responsible AI or for the west where our history has shown that we’re failed to be responsible and even the current reality shows that we fail to be responsible yeah go Happ yeah absolutely that’s a fantastic question and thank
You for bringing us and reminding us of the reality that has been suffered by many many people people right here right now uh I heard somebody say um that it it’ll be okay be okay when AI aligns to human values and then it’ll all be great
Um I think it was Tom who told me about this actually one of our leadership team and you know if you pause for a moment and you think about the Millennium that has gone past and the the misalignment of many values and the infliction of suffering by peoples on other people
Then to assume that AI is suddenly going to come out of nowhere and align and become Our Savior well it’s that’s just idolatry I think now just because a country like uh the UK has had its problems in the past and will have problems in the future doesn’t mean that
We shouldn’t be convening a conversation uh and of course what we’re doing in RI UK we’re not government we’re not part of that mechanism we’ve been funded by taxpayers to open the comp conversation up and to hear you know absolutely those kind of challenging voices as we go
Forward got no I’ve got no kind of feelings that we’re going to solve everything but the more people that can critique Challenge and have an equal voice the better can I also add to that um people do ask me you know how do we become a well leading research program
Why why do we claim this and and I like to say that we we don’t try and claim that we we’re going to solve this for the world in the UK and I can talk about the the projects we’ve just funded four or five of our projects are with
International Partners we have a project that’s just been funded and the list of of funded projects will be on the website soon one is with Japan looking at emotional AI the issues that arise when you deploy AI as companions another project will be looking at AI uh issues
And regulation across the UK and the EU so working with Partners in the EU another project is going to look at um AI adoption bymes across across Africa um and with India as well we have an upcoming workshop with with India so we’re not trying to claim that we are
Going to solve this on our own we are actually going to get everyone to work together on this and one of the plans we have with the launcha partner Network which will will come out soon as well uh is to bring in companies from across the
World not just big ones but small ones as well uh Civil Society um to this to this to this conversation to bring up some of the issues that you’re concerned about um and uh one big ambition I have is to to build an alliance of academic um institutes
That across the world irrespective of of of of States whether and and in China being part of that conversation as well to ensure that all these voices are heard and and and we build a eye that works for everyone not just not just for the UK and that tries to expound on on
On British values thank you great there’s someone here had their hand up very quickly in the middle of gentleman in the white shirt hi yeah thank you uh Luke christopherus from Barkley’s Eagle Labs um Chris you may have touched on this um uh earlier or or maybe not and maybe I
Misunderstood you but it got me thinking is is there a prisoners dilemma with AI research are there states that maybe are you know fostering irresponsible AI research and if that if they are what would does that look like could you contextualize and maybe give that some
Texture what would the risks of that be so so the UK and the United States in particular have agreed um on some or or they have very very consistent ethical guidelines for the use of machine learning in military applications and they include things like proportionality which is that you
Shouldn’t you shouldn’t use Force beyond the intended effect it should be absolutely minimal um human centeredness is another issue and and if you take human centeredness in in many existing applications they’re not human- centered so from the perspective of machine learning so there’s a requirement on people thinking about putting machine
Learning into military systems that the human should remain in control throughout most of the things that might be used in Ukraine and elsewhere that’s not true once you once you launch them you don’t have control over them so they put constraints which mean that people actively don’t do things today in the UK
And United States which is technically feasible to do in fact technically not trivial but but relatively straightforward to do but we’ve decided and this partly the answer to the previous question there that we we should not allow that technology to progress at the moment I’m 100% convinced that other countries around
The world and some of them have been mentioned previously don’t do that and they they have active research projects for which uh those constraints that we accept and impose on ourselves are not considered relevant that make sense can I also add to that that you know in terms of uh
Very quickly um the technology and and some people are going around saying oh we should Reserve this technology to few companies that is not true when it comes to ai ai is code and people are working every day in the academic Community to make this code more efficient and you have people
At synon University Building CH GPT um equivalent solutions that will run on a laptop right and similarly at Stanford so these Technologies going to get more accessible to more people around the world and it doesn’t make sense to just say oh we just reserve it to some
Companies that only that that’s a that’s a false sort of argument to use thank you great got a question from lady at the front hi thank you uh my name is rasham I’m Global head of policy at the open data Institute we’re not for-profit looking at an open data ecosystem and obviously
AI is a big part of that at the moment um one of the challenges I think we see and it was touched upon earlier today um was the scale and pace of the AI research that comes out um mostly funded by big Tech and when I look at civil
Society and you look at the policy teams and the research teams that are kind of often working on a shoestring budget they just don’t have the resources the people the funding or the scale to be able to respond at the speed and Pace at which the tech sector is driving that
Research so I’d love to hear your thoughts on how we Foster responsible AI research that’s coming from civil society as well not just Academia and big Tech both of which are important but we’re not seeing a holistic ecosystem when it comes to that research because that scale has increased so quickly
Civil Society can’t keep thank you anyone want to take that I I I I can answer that in some ways um give some time to the my my panelist panelist to to to think of another answer but what I found is that we are starting to see uh some some Echo Chambers forming
Around ecosystems so some people just wanting to talk about AI safety so they bring people that think the same way and and and and have uh can can can we’ll just build on on the same story then you have you know uh maybe social scientists who will think about you know immediate
Issues immediate threats and and and issues of bias and and they and and and equally on the technical side and I come from an AI background I’m part of this academic community that don’t talk to the social scientists so I think there’s a there’s a there’s a there’s a divide
Here that needs to be to be to be bridged um we need to have people from the social sciences from the soci tech soci technical communities coming to these conferences most of the technology we say that that are being built by these big Labs is actually published at
Those conferences I see them there but their understanding of ethics of safe AI is very different from what other disciplines would consider safe Ai and there’s no value for for people to come together to talk about these things when we bring in the experts from from those
Domains uh together we are trying to fix that with responsible auk we’ve seen this problem in the trustworthy autonomous systems program we trying to fix fix those those conversations great just quickly you go Ahad I think we need a civil service reform act um I think we need to totally change
The way that we fund and employe civil servants um I think that the degree of technical skills and Technical competency is not what we need for the next 20 years and and I think that there is a an over an overabundance of prance of social scientists and political economists and an absence of
Technologists and and so specifically things like the continual rot of civil servants in key technical jobs uh has fundamentally weakened this country for The Last 5 Years and there’s a reason why we struggle to keep up with technology in terms of our policy and that’s because there’s not enough technologists in
Government and I think you know I think there’s some some things those about ukri I think some things that ukri have done which are absolutely brilliant which is uh public policy fellows so getting scientists and technologists to go in but go science needs to change and
The CSA role needs to change and there absolutely fantastic opportunities there to strengthen this country and to make it more confident in our public policy can can I end with a very practical answer as well so ukri also funds something called includ plus.org and that will fund directly Civic Society
Organizations not directly just academic so please get involved and start the conversation thank you this is a very good point sorry we don’t have any time for more questions because we’re going to hear from the boss of UK so we have to move off the stage for um oine but
Before we do just want to say thank you and also just one word people here want to get involved with what you’re doing how do they do that because that’s fundamental thanks M um if you want to get involved with our auk sign up to our mailing list we’ll be announcing a big
Funding call 10 million pounds for um three large research projects multidisiplinary research projects think looking at responsible human AI organizations AI um looking at ai tradeoffs ai harms and AI law and regulation so uh sign up the mailing l is join our partner network if you’re an
Or a large uh business and uh and watch out for our projects we will be uh announcing them and uh you are all welcome to collaborate with these projects and and and jump in and and and bring us to your events um so we can come and talk about the results that we
Are producing great I don’t know about you I always feel reassured when you have thoughtful people thinking hard about really complicated things who are not politicians so thank you for the work that you do and thank you for the panel today thank you Mar [Applause]