SPEAKER_03 0:12

You're listening to the last podcast with broadcasts. Hello, you're listening to the Philanthropisms Podcast. This is the podcast where we try to put philanthropy in context. I'm your host, as ever, Rodri Davis, and this week I'm joined by two guests. I'm joined by Milos Maracic and Giuseppe Ugazio. Now, Milos and Giuseppe are the editors of a new volume, the Routledge Handbook of Artificial Intelligence and Philanthropy that's just come out, which I should say, full disclosure, I have a chapter in, and more of that in a bit. But I was delighted to be joined by Milos and Giuseppe to talk about the new volume and also just sort of more broadly about the whole question of the intersection of AI and philanthropy, which is something I'm very interested in. And I attended a conference that Milos and Giuseppe organized earlier this year in March in the University of Geneva, which was linked to the book as well, and we'd sort of touch a little bit on that in the conversation. So just a bit of background before I sort of tell you what uh to expect in the podcast. Um so, as well as being uh editors of this volume, um so Milos himself uh is an expert on AI and an entrepreneur. Um he's also the founder of uh the altruist league, which is a global network of sort of ultra-high net worth people and foundations um that uses AI uh to try and kind of make their approach to philanthropy as effective as possible. And Giuseppe uh is the Edmund De Rothschild Assistant Professor in Behavioural Philanthropy at the Geneva Finance Research Institute, and he also leads a project on artificial intelligence and philanthropy at the University of Geneva as well, which has been running for a little uh while now and is how the um uh the Routledge Handbook came about as a as a publication. Um so I sat down with Giuseppe and uh Milos uh fairly recently, and we had a really good, wide-ranging conversation. So, as I say, we kind of talked a bit about the the Routledge Handbook, how it came about, what the aim was, what some of the content, and how they sort of decided on the themes within it. Um but we also talked kind of more broadly about AI and philanthropy. So we talked a bit about what are some of the interesting and exciting examples of AI currently being used um directly to try and kind of achieve uh social or environmental missions and how we should be thinking about those. Um we also talked about how AI could be used within philanthropy itself, so both to kind of understand philanthropy more effectively and to map it and to kind of be able to spot patterns and make predictions potentially, but also how we could deploy AI in various ways to make philanthropy, you know, quote unquote, better or more effective, um, and what some of the opportunities were there and some of the challenges potentially. And we also talked about what some of the barriers are currently to getting um philanthropy to use and adopt AI effectively, some of those barriers in terms of the availability and the quality of data, you know, the lack of knowledge about what AI is and how to use it, um, the difficulty often of having the resources, uh, both in terms of financial resources and time to invest in this, and whether that means we need to put more focus on partnerships between the not-for-profit sector and the for-profit sector, and and what some of the challenges uh in that might be. And we also talked about some of the kind of known challenges with AI as a technology around the way it can be used for misinformation and disinformation, uh, its tendency towards uh creating or exacerbating bias that already exists in data and also the challenges that it might present to kind of identity and trust in the online environment. So lots of uh subjects covered. Um so without further ado, let's go into the conversation. I will be back at the end for the usual bit in housekeeping and tidying up. Uh, I'll also, I think, add a little bit where I'll just talk briefly about my particular chapter in the Routledge Handbook and just give you a bit of an overview about that in case you're interested, and then you can go away and read it for yourselves. Um, so I will be back at the end uh to do that. Okay, great. Well, I'm here with Milos Marichic and Giuseppe Ugarzio. Hi to both of you. Hey Rodri, good to be here.

SPEAKER_02 4:53

Hi, Rodri.

SPEAKER_03 4:54

Yeah, well, great, great to have a chance to talk to both of you. I mean, we've kind of been in the same room uh before at various times and and talked quite a bit about the top, well, various topics, but the topic of philanthropy and AI, which is what we're here to talk about today. And I guess the starting point is the you know, the hook for this is that you have both co-edited um a new volume that's come out, the Routledge Handbook of Artificial Intelligence and Philanthropy, which is a big volume with loads of um chapters in it that have been contributed by all sorts of authors, covering lots of fascinating uh insights and perspectives on AI and philanthropy. So maybe the best starting point is just for you guys to say a bit in your own words about why you were interested in pulling this book together, what the thinking behind it was, and what the process was, and sort of how you went about structuring it.

SPEAKER_02 5:40

Thanks a lot. Thanks a lot, Roger. Thanks for the opportunity to talk about this uh this uh book and the the research project uh more broadly. So the original uh thought of putting this book together was seeing how much uh AI was uh sort of catching up. That uh the idea for the book came up before uh all the big explosion with uh ChatGPT and so on that put basically AI in everybody's mouth. Um but it was really realizing how little awareness uh there was in the philanthropic sector um about AI. So you had the the niche um the niche actors and Millis and uh his organization being one of them that were really savvy, really knew a lot about AI and really exploited the potential of AI a lot. But then if you turn your head to the rest of 99% of the sector, uh you would see that you know they they were they were very either scared, not ever uh not ever, you know, not even you know thinking about it, they never thought about it, they um you know they they seldomly uh even uh realize you know what how they could benefit from it, right? So uh we thought, Millis, you have a lot of experience, you've been using it forever, you've been advising philanthropists thanks to AI. From our side, we've been researching how to use AI to build tools that could help uh philanthropists, say mapping different actors. That's uh one of the projects we have. Uh, why don't we uh get together and uh you know bring uh experts in this uh in this in these two fields um to to you know start a dialogue and try to bring them back together, right? If you think back at the origins of AI, a lot of um the first uh movers were actually philanthropic organizations, right? But then it didn't really explode, it really stayed in uh in this niche. So then uh the the idea was okay, let's try to uh to do some uh capacity building, spread a bit of knowledge, show people how AI is being used in philanthropy so that maybe more will be interested and then uh uh catch on to it.

SPEAKER_03 7:48

And and Milos, yeah, what's your kind of ambition for what people reading the book might take out of it? You know, philanthropy practitioners who are interested in this topic, they come to the book, what can they what can they learn? What can they get out of it?

SPEAKER_01 8:01

I guess what's uh really fascinating for me is how how b bidirectional actually the relationship is. So uh there is one aspect of the relationship that has already been established. We've seen uh AI used in philanthropy for many years now. Let's say on the on the nonprofit side for finding better donors and maintaining relationship with them, communication and so on, then on the side of the donors, uh calculating risk, uh, calculating impact, uh, doing better donor grantee matching. So this is something that uh we are increasingly used to seeing, but there is that uh other direction that I think the book is probably the most fascinating aspect of the book for me, which is what can philanthropy do for uh for AI. And that's maybe something that I might like to accentuate because uh philanthropy has this fantastic opportunity to be a neutral convener. We are all parts of different AI debates, we know how how difficult they can be and how many different points of view philanthropy has a role to play there, and then also um simply by funding organizations that uh inform citizens about their rights with regard to AI, that uh that you know protect uh protect the public interest for lack of a lack of a better expression. So uh the I think it's only when the book um when the book was completed and when I actually read the the final the final version of it that I came to understand uh that that you know you know twofold bi-directional relationship. And I th but now I see it as probably the the strongest uh selling point, so to speak, of the book.

SPEAKER_03 9:37

Yeah, absolutely. And I think that you're right, there's that, I mean bi-directional, but maybe even tri-directional, in that I sort of look at it and there's you're absolutely right, there's that bit which is about the way in which philanthropic organizations or nonprofits are using uh AI, and then there's the bit about how they are potentially informing wider debates and influencing the the direction that AI is being developed. And maybe there's a bit in the middle, which is how are philanthropic organizations themselves being affected by AI, even if they're not the ones driving the implementation of it. And I think the book covers all of those different things. Um I'd love to pick up on on all of those. I guess maybe let's take the the sort of um the optimistic uh utopian bit first, which is often the bit that's kind of front and center. What are some of the examples of non-profits involved in applying AI tools directly to kind of their cause areas that you found most interesting and exciting?

SPEAKER_01 10:39

I think uh what what um I think the the uh results are in, so to speak, uh uh and they have been in for for the last uh couple of years. And I think we're at the stage at which it's no it's no longer a matter of talking about you know individual examples, but we actually have statistics from from the UK, from the US that are showing that that um NGOs that use uh AI. Um and there are there's a there's a penalty of platforms uh out there that that that propose that uh use AI or purport to use AI to help nonprofits find better donors. I mean there there are there are very conclusive statistics by now that are showing that uh these organizations are spending you know up to 30-40 percent less time to to capture um to uh to capture up to 20 to 30 percent more uh donations uh in in some cases. And then in terms of uh donor churn, uh the statistics are also looking good in terms of um retention, in terms of you know that um improvement on communication, so increasing the number of touch points and so on and so forth. So I think what what's actually the most encouraging for me is that it's not no longer a matter of saying, hey, you know, I know this organization that's using AI and they're you know they're they're doing great, but it's more of a it's becoming a massive trend where if you're not using AI, you're you're missing out. That's the way maybe I would frame it.

SPEAKER_02 12:03

Maybe to build on uh Milos what you're just saying and bring an example or two examples from the book, uh we have this chapter by uh Marta Herrero and uh Shona Conkanan, uh, where they discuss how you can uh enhance uh basically fundraising thanks to to AI, right? And specifically they look at uh um basically uh digital method, the digital methods for fundraising, which is of course where AI has the largest potential, can really be used to you know kind of uh get a deeper understanding of your donors or your donor base and then uh build sort of a storytelling line that is sort of more appealing to them, right? That's sort of uh with for the fundraising, and then in a way to sort of uh bring uh actors together and try to coordinate efforts, which is one of the themes also in our forward uh by Professor Floridi, where he's actually seeing uh one of the largest potential for AI is really to help uh people collaborate more together, organizations collaborate to achieve the very important aims, for example, of you know, addressing some of the SDGs or contributing towards some of the sustainable development goals, right? So that's where you need uh concerted concerted efforts uh to really make a difference. You need a systemic approach where you can bring the different people together and map how they uh contribute, right? And then we have uh uh this chapter by our uh friends uh from the uh let me just get the right name so don't uh make it run. It's Catalyst Balkans, uh, where they use uh large language models, right, to find to map uh in the in this uh European region uh who's working on what and produce then a taxonomy of the philanthropic organizations there and the initiatives going on. And this also, I guess, resonates with what Millis is doing with the altruist leap, right? Of finding uh people uh and organizations and match them based on interest or moral values or or or other uh dimensions. And maybe Rodri also uh brings uh back to mind your chapter, right, on uh how you can uh use AI then to uh look at the individual donors, right? And uh you you can sort of uh uh get a more comprehensive picture of that to understand the, if I remember correctly, the how, when, and where uh they will want to get uh engaged.

SPEAKER_03 14:23

Yeah, absolutely. And I I should say I might at the end uh after we finish speaking on this episode give people a sense of the detail in my chapter. But it's it's essentially, as you say, about the way in which AI and particularly kind of algorithmic recommendations of various kinds can be used to inform the choices people make when when giving to organizations or causes and some of the ethical issues that there might be in doing that. Because I think there are huge opportunities, but as with so many things with AI, if we rush into doing them without sort of prior thought about some of the unintended consequences, there are maybe causes for concern. You mentioned there, Giuseppe, the the example from the the Balkans. And one of the things I think is really interesting about this volume as a whole is that there are lots of different examples and perspectives from different places around the world. You know, there's a really interesting um chapter, for instance, on China and the way in which the sort of speed of adoption um uh because of the the um rate of of uh growth of digital giving in China has has been pretty enormous. Um, I wondered whether you had any sense from from this work or your sort of wider experience of whether currently adoption of AI among nonprofits and philanthropic organizations is particularly concentrated geographically, you know, is it distributed well or are there places that are further ahead or further behind? And also in terms of cause areas, you know, is is it that the areas in which there is a lot of data already are the ones that benefit? And are there other areas that are maybe being left behind?

SPEAKER_02 15:52

Yeah, absolutely. So basically you're touching on uh one of the biggest risks uh or biggest uh downsides of using AI for everything now, which is basically uh exponentially increasing the digital divide, right? If you don't have data, you don't you can't have any answer from AI. AI doesn't know a certain problem, a certain uh issue exists. And then if people use AI, say to find causes they want to give to or organizations, then you have no data on it, then you'll have a big issue, right? But to come back to your point, so I think it goes really hand in hand with the how much a society has been pushing uh for uh AI. And here basically you then would see uh China and the US, you know, with the trying to really have the dominant uh LLM uh that everybody is gonna is gonna use, right? And uh they then you see the uh some of the largest users, users of AI are uh, for example, the or promoters of AI, right? The McGovern Foundation is US-based, and we actually have one chapter also from them uh in in the book. And as you mentioned, right, uh China has been the Ten Cent Foundation, so under you they've been using uh AI massively, uh, even and you know paired with uh TikTok and uh this social media to sort of reach uh uh unprecedented uh levels of uh fundraising from individuals. Uh so I think that when a society engages a lot in uh promoting AI and digital technologies, then also the philanthropic organizations will sort of follow suit and also be prepared. So maybe one of the next uh regions that we will see uh catching up on this is the UAE, the MINA region, just because how much they're investing already in finance and in other digital technologies. And but hopefully, also it will um one of the hopes is that when philanthropy uh realizes and tries to combat, say, the digital divide or these issues, then we will also manage to uh see that regions uh like the bigger African continent can leapfrog and start uh you know getting more advanced and digital and then um being digitized and then using AI to help their um you know addressing their issues, right? And we know there is uh quite a big network of local uh entrepreneurs that are uh leveraging AI and trying to bring AI to uh to to promote uh growth also in that region.

SPEAKER_01 18:14

Yeah, and and just to build exactly on what um Giuseppe was saying, I think a lot of this is is um yes, society driven, but also donor-driven. And uh, for example, when we were starting out close to 10 years ago, you know, our first clients were these uh tech sevy uh founders in in California who got the whole story around AI, we're very excited about it, and were frankly happy to pay our prices, which were you know 10 times higher than the price of an average philanthropy consultancy. And so, you know, through through collaboration with them, we we started to uh to deploy our first models and and to get some traction. And I think this is now spreading around the world, driven by the donors themselves. So there's a certain uh section of donors that are very tax-heavy, um, they you know they're excited about AI, they're also a little bit potentially disillusioned with the with a certain version of philanthropy that they're seeing. So they want things to be a little bit more systemic, a little bit more um trust-driven, uh trust-based, and so on. And these people are now, which I find really beautiful, appearing all over the place. They're appearing in Africa, they're appear, they're appearing in East Asia, like like you both mentioned, uh, you know, there's an explosion of such people over there. So the trend is in that respect becoming global. But then for that to trickle down to the to the local level, uh exactly like Giuseppe said, it will take um, you know, it will take uh efforts of uh of uh different sectors, including governments, including you know institutes, uh think tanks, and then and so on, uh, for that to be comprehensive.

SPEAKER_03 19:54

Just to build on your point there, Mirosh, is is the flip side of that, and I guess one of the maybe the concerns is that there is obviously this huge opportunity around AI, and as you say, there are donors potentially that are going to drive that because they have the existing experience and knowledge of these tools and they know that the potential they have. Is there a danger sometimes if they aren't able to make the right connections with the organizations on the ground that know what the genuine needs and use cases are, that you get this kind of tech solutionism problem where people come in with it with you know the idea, I want to use AI, rather than starting from what is the actual problem and can this tool be used to to address it? Have you seen that happen at all? And how can we kind of mitigate against the risk of that happening?

SPEAKER_01 20:38

I've seen it happen uh quite a lot, actually. And I do think that it's uh it's a big problem. And a lot of it has to do with the the work of the intermediaries between the donors and the and the causes on the ground. So if we're talking about the work with the donors, then uh there is something to be said about donor education, so helping them along their journey. They do come with preconceptions, with ideas about what they want to do, but then it's uh our job, I suppose, to talk to them, to to to try and shape, shape their intention in in ways that that uh can can produce uh impact uh on the ground. But the work is also uh there's work to be done also on the other side, because uh obviously one of the biggest problems in in creating systemic change is that we simply don't have uh a perfect overview of all the causes on the ground uh that are worth investing into. So that's something that the Altrist League works invests quite a lot of money and time uh into. And so it's uh we also say it is our job to to first of all find those causes, to track them, to to pre-select, you know, you know, the the best ones, the most impactful ones, the ones that can use uh funds the most um easily um and then to make to make the matching between the two, all the while maintaining that uh may maintaining a certain balance in the conversation. So it's not the donor coming in and saying, Hey, we want you to do this. And uh here's the money. And it's not the it's not the nonprofit either saying this is what we do, either in either invest or or leave us alone, but you know, to make sure there is some sort of a space there where there can be a little bit of um discussion to to find some mutual ground for because our job is to make sure that the collaboration ends up being not just one project, uh one you know initiative, but a long-term trust-based investment over five, 10 years and so and so on.

SPEAKER_02 22:29

Maybe to add up that, uh, one of the risks, uh the like broader risks. So these are really concrete examples that Milos share, but in general, is not only for philanthropy, right? But it's for any uh use of the eye when we use it to substitute instead of complement, right? If you're asking AI which cause should I support, is you know, it's really the wrong answer at the wrong question, right? Because then you get answers that will be biased that basically just reflect the same or even amplify, right? The same uh the same bias, like searching for for uh you know a missing object under the light and not where the dark is, right? The AI will uh just point at different points in the light, but then what you want to do is actually find out up in the dark places where should I point out my flashlight to find these better solutions? And that isn't probably not what AI can do. AI can tell you maybe you know what type of light you should be using or you know which uh which angle to take or how to, you know, once you have uh identified partners, what could be some sort of good ways to interact or to develop uh which tools you could use to make sure the the uh the interaction is efficient. Uh, but it's really this uh the broader issue of substitution uh that sometimes you know we when we become lazy and you know, if we the same, when you answer an email, if you ask AI to answer the email rather than just polish your English, right? Or at least not yours, Roger, but mine, and that that's what you can use AI for. You shouldn't really use it for substituting yourself.

SPEAKER_03 23:55

And and in the context of the possibility of using AI to apply it to sort of philanthropy as a system rather than specific cause areas and kind of make the allocation of resources more effective. As you say there, Giuseppe, if you just did that without giving sort of appropriate thought, first of all, the danger would be that you would focus, you know, it would kind of reinforce existing biases, you'd only know those things that you know, and it wouldn't be able to kind of tell you anything outside of that sphere of knowledge. But if we want to be able to do something more interesting with it, I guess the starting point is do we need then the data on philanthropy at a at a kind of more universal level rather than in in silos? And how far are we off making that a reality? I mean, what what more is there to be done to create a data set that allows us to do something meaningful at a kind of systemic level with philanthropy as a whole?

SPEAKER_02 24:48

Yeah, so unfortunately, I think we're very far. And I think there is an endemic issue with uh philanthropy. Uh and the first one is there is very uh a big lack of transparency, right? What the information you get, first of all, is only success stories, right? You try to find an annual report or a project report from a foundation or an NGO that said we failed. Uh so you know that that's the one side, right? And then what they what you see in these reports are sort of the tip of the iceberg is like the most appealing uh projects that then you know bring you uh good visibility and so on, but then you don't really have the exact data uh or the concrete data that you can then maybe share so to prevent someone else from failing, you know, they already see that your organization took this approach and it didn't work, and then either find out by failing myself, or it'd be great to see this, right? Um so unfortunately at the moment we're quite far, but I think the more we see the benefits, for example, from science, from open science of uh sharing data, and there are safe ways to do it, right? It doesn't just mean throw everything out there and let everybody use everything. It means you know, there are ways to share data in an appropriate way so that it can be uh you know very efficient and um and constructive. So when we do it properly, then in other fields, hopefully financially will see as well uh the benefits of um uh of sharing data. And I've I hope we will move in that direction. And to this end, we are uh collaborating with uh Philea, this uh umbrella organization of philanthropic organizations to promote uh data commons initiative at the upcoming uh conferences. This European research network on philanthropy will be next year in September. And we'll try to have a conversation with researchers and with uh philanthropists on how and what we need to build a data commons for philanthropy to make it, you know, uh that it's anonymized, it's uh uh safe, it's stored in uh in appropriately, and so on. So that then we can hopefully start with a few volunteers, organizations sharing the data and maybe generating. Look, this is a report of what has been working, not working, how much funds have been given, and this these are there that people are not working on, and it's only based on one percent of organizations. Imagine if we had 90% of organizations, what a much more richer, uh much more rich uh picture we could get, right? But uh unfortunately still at the very early stages.

SPEAKER_01 27:13

And I don't know, Midas, if uh in your experience you've seen Yeah, maybe maybe two things to add, uh plus a bit of a big picture um overview. I guess uh yeah, one striking thing when we work with donors is always uh how uh um different what they uh actually do uh ends up being as compared to what they say that they do for not, you know I I don't mean to say it in a bad way, but it's just some something that seems to be a practice around around philanthropy. So so the the public persona is one thing, and then you actually, after a lot of trust building, actually look at their um investment history, and it it it shows you a slightly different picture, which is either like Giuseppe said, slightly less successful or perhaps slightly less coherent and so on. But it's it's at least the big the beginning of a discussion. On the side of the uh on the uh the data set of uh let's say NGOs or citizen movements and so on, um, there is an absolute shortage of that. And there, again, maybe as an entrepreneur, I see I see a problem because uh, if I think about my my company, that is probably the the core of our added value, right? So it would be very difficult for somebody in my position or similar similar organizations out there to simply share that uh indiscriminately. But then like uh like Giuseppe was saying, there are ways of sharing that um that uh can be a win-win, because you can mask data, you can you know share some as some parts of the data set um and so on. So there are there are there there are definite opportunities for uh for collaboration there. And uh the last the last component of of of um what I would like to say um is a slightly maybe slightly different perspective because your your first uh first sentence in in this question, Rodri, was uh uh about general AI use in philanthropy and where it's headed. And uh for me, um it's quite simple. I I don't think uh that we can afford not to use uh AI in philanthropy in the future. And the reason is uh very simple. We are uh working with um uh increasingly complex data sets. So we are working with uh donors that have the whose investment history spans uh thousands of organizations, where we are working with a data set of potential candidates which might have hundreds of thousands of of um of organizations in there, and so somebody needs to make a matching between the two, a human can no longer do it. And uh so it's a it's a simple necessity. And then if you throw uh throw additional components that add complexity, so you're trying to be trust-based, which means you want less reporting, you want less you know to bother organizations less, and you want to leave them more to do their own work, and uh then you um do your impact calculations and tracking and so on in the real world as opposed to through the through the um uh reports. Well, then you need yes, analysts uh to do some of that manual work, but you also need AI, you know, through all the different scraping techniques and NLP to calculate this. And then maybe the last component that adds even more complexity is we're all increasing increasingly talking about systemic change. So not just doing one intervention, but actually uh working for the for you know for the long term um with the different layers of the society, from think tanks to you know village elders to social organizations and so on. So having that broad front approach, well, that adds complexity because you need to now model the actions of all those uh all those uh actors. And in addition, maybe it's the last component, uh philanthropy sometimes has the tendency to act as if it was acting in a vacuum. So a funder comes into a country and says, Well, this is my operation, this is what we're trying to do, these are the KPIs. Fortunately or unfortunately, more often than not, you are just one actor in uh you know, in an environment of many different ones, both other philanthropic organizations, government entities, uh social organizations, and so on. Again, something that you need to understand, model, predict, and so on. So to me, it's very simple. AI just has to be used in philanthropy going forward. It's just a matter of how we do it responsibly and effectively.

SPEAKER_03 31:40

Yeah, I like what you were saying at the end, Milos, because I often say one of the most sort of pervasive and dangerous bits of thinking in the whole world of philanthropy is the the myth of the lone saviour. It's this idea that actually all the best change comes about because one brilliant individual or organization comes in, thinks of something nobody's ever thought of before, and kind of changes everything. And that almost never actually is the case in terms of how change happens. But I think it's such a kind of, as I say, a powerful um idea that it's it's sort of stuck around for a long time. I wanted to ask actually a bit about the role of the private sector and particularly the sort of tech sector in all of this. Because I guess one of the things that I'm thinking is that if we're talking about AI usage and adoption at scale across the world of philanthropy, the reality is that certainly when it comes to designing new tools or kind of applying these tools to new areas, very few nonprofit organizations are going to have the knowledge and skills and resources to do that themselves. So they're going to be reliant on partnerships with the tech sector or the for-profit sector. What, you know, what do you think the potential is there to bring together the sort of profit motive within this and and some of the interest of nonprofits? And do you think there are kind of potential challenges also to be overcome there because there are going to be power differentials between the different organizations and maybe misalignment of you know values and objectives?

SPEAKER_01 33:06

Yeah, maybe to preface this with uh to just just saying that I um another hat that I wear is uh as the chair of an industry organization of institutional investors, which is uh quite active in in all the regulation discussions in in Brussels and and so on. So I have also that uh very much for-profit uh perspective. I think uh the you raise a very valid point, and I think uh philanthropy has a role to play there. Like I was saying earlier, as a neutral convener, philanthropy has that uh luxury or an opportunity to be on the good side of everybody. So to understand the motivation of the tech entrepreneurs, to understand the motivation of the government, of the think tanks, of the of the civil society, of the ethicists, and so on, and just to bring them all together to at least have a uh a reasonable, honest discussion. And I think this is um this is a great uh great beginning. Then uh when it comes to the role of the tech sector in particular, so to answer your question directly, the the answer is twofold. First of all, a lot of these uh tech entrepreneurs are also philanthropists, as I was saying earlier. So they uh they also do at uh at some level have an interest in in this technology being developed in a way that is that is ethical, that benefits everybody, and they uh I think they stake uh a certain percentage of their identity, at least those that I that I speak to, on on that. They want to be the people that make that happen. Uh but with all that said, uh regulation cannot be obviously left to them. So I'm a huge believer in regulation. It's uh it's a very m maybe mundane and an unsexy way to look at things, but we absolutely uh we absolutely need the uh need the regulation uh to to to basically you know limit uh to assign responsibility where it should belong, first of all, and then you know to limit the the potential negative effects uh of uh of the technology, in particular on those with the least power, so so marginalized groups and and just individual citizens.

SPEAKER_02 35:09

I mean to bring back also to to the original question on you know what can philanthropic organizations, especially NGOs, do when they don't have the sort the resources, right? It's so first there is always the hopefully uh that the donor organizations can already think of including and say an AI development, or uh like our colleague Francesca Bosco from Cyber Peace Institute always uh tries to argue for, you know, that you should also include some sort of cyber readiness or some uh infrastructure uh part in the grants, right? That we have overheads for people probably we have overheads maybe also need for the infrastructure and the digital part uh as well. But then I don't think this is gonna be such a big concern, right? Because the tools that philanthropy needs, they're tools that other industries need as well, right? Impact measurement, you need it in finance, you need it in uh and you need it in philanthropy. Uh HR management, you need it in any uh basically large issue and so on, right? So I don't think there is gonna be too much of a need for tailor-made tools for philanthropy that are not already developed in uh sort of by the market and then can be sort of uh brought in to philanthropy as well. And maybe the one thing that philanthropy can actually, from the donor side, try to promote is uh you know trying not to have to change everything to fit uh the digital world, the AI world, because it would basically mean reinvent every single philanthropic organization, and maybe at the NGO level, at the communities level, that's really impossible to do, and actually promote AI that is flexible enough. The problem shouldn't be that it's me not using properly the software, the problem is the software is not adapted to me, right? It's uh uh I shouldn't be finding way around to trick the software and still use that this AI tool when I can't, uh, but rather uh because it's really a rigid tool, uh, but uh it should actually be a more uh on the AI developer side to find something that fits the existing world so that we don't we don't have to change everything we do.

SPEAKER_01 37:12

Yeah, and just to just to maybe reinforce the point that uh Giuseppe just raised uh around uh the similarities between the for-profit and non-profit sectors. So when we were when we were just starting out uh in the altruist league, we we uh worked a lot with the for-profit sector as well, just to just to pay the bills because they were they were way more receptive. And it was it was just amazing to see how the challenges were um almost identical in terms of in terms of you know team building, understanding the technology, understanding uh what organizational changes are necessary around it so as to deploy the most effectively. So there are huge opportunities there. 90% of what is done in in the for-profit space can be with some modifications uh used in philanthropy. Yeah, absolutely.

SPEAKER_03 37:56

Um I just wanted to pick up on the question around um sort of law and regul uh regulation as mechanisms for shaping the development of of AI. Because we mentioned at the start of the conversation that you know one of the points of relationship is um between philanthropy and AI is the role that philanthropic organizations play in kind of informing that that debate. And you touched on it in your answer before there, Milos, that actually maybe a couple of years ago, the debate around this focused on ideas like making AI more ethical and sort of relying on individuals and organizations to do the right thing, and that maybe that's shifted more towards recognising that even among well-intentioned actors, you just need rules and and laws often to kind of govern things. What do you think is kind of the areas that are most of relevance and of interest for philanthropic organizations and civil society organizations, particularly in Europe? Because I think I I got the sense at the the event that we uh were all at in in March at the University of Geneva, there was a lot of discussion there about how actually Europe feels as though it's maybe slightly further ahead when it comes to questions around regulating AI and actually taking a relatively tough stance on it. So I wondered if you kind of had any thoughts on how involved philanthropic organizations have been in that and you know where they had had brought influence to bear.

SPEAKER_01 39:15

Well, first of all, maybe to answer the the first part of your question, which was around the um around the regulation and and um and how how philanthropic organizations should um approach it. The um at the EU level, the regulation is now in it's uh in a very interesting phase, I I think. Uh so the law is uh so to speak in in place, but the standardization process uh for the law is uh is ongoing. And I find that that phase to be crucial because that is exactly where you will define what the legislation really means. So when you say that you need a risk management framework, what what what does that actually mean in reality? Um and similarly for for other reporting requirements. So my my first piece of advice would be for philanthropy simply to be involved, first of all, to understand what is going on. It is it is difficult enough to understand what the what is going on in Brussels uh for us who are following it full-time, but you know, I would still make an effort uh because there is uh there is an opportunity in many of these bodies for foundations to be to be simply involved in day-to-day, day-to-day discussions, and some of them potentially represent actually the uh the interests of the broader industry. I know at least our our um our the friend of both of us, uh Sebastian Hollensleben, who chairs uh JTC21 and whom you know as well. You know, he's he's extremely welcoming to these external perspectives. And uh and then uh we are seeing philanthropy, some philanthropic organizations already involved in creating fora for this discussion. For example, Giuseppe and I will be in Paris in a couple of weeks for an event organized by the Bertelsmann Stiftung in collaboration with the OECD. And it's uh the topic is going to be digital trust and digital identity. Things like uh, how do you know that that the person who wrote something or the who with whom you're speaking online is is a is a human and not uh not a bot. Um so that's an example of um philanthropy just uh taking an initiative, collaborating very importantly with another organization and creating a forum for for uh you know some exchange. And then the third the third um area, as I touched on previously, is uh simply funding organizations who are doing good work in the in the field, protecting the the rights of the citizens and informing them. So organizations like Algorithm Watch and so on. So that's uh that's maybe uh the lowest effort, but potentially very impactful opportunity. If you don't want to be part of all these discussions, you don't, you know, you don't have necessarily the time to attend the fora. Um you can you can direct your funding towards organizations that contribute.

SPEAKER_02 41:50

Yeah, and maybe the other one just uh to complete the the picture is uh for advocacy, right? We've seen it in uh uh for any uh any important topic that's uh philanthropic organization NGOs have been uh quite vocal and trying to promote more legislation, right? And so that's the the other way in which philanthropy can get engaged also into that direction.

SPEAKER_03 42:12

And and do you think because you you mentioned the question there around um digital identity and and sort of trust and online authenticity, which always strikes me as one of, you know, if you're trying to convince an organization within civil society that it should care about AI when it's not, you know, already minded to, I often find if you you know start talking about the bad stuff maybe first. And the one that tends to cut through is that, because actually there's so many organizations trust that, you know, the trust that the public and their supporters have in them and their ability to, you know, have authenticity and to have people kind of respect and trust what they say is so vital that when you point out that there are these new threats to that that might you know lead that to be undermined, they suddenly realize that it is very relevant to them. Do you think that that's one example of how we need to sort of frame some of this debate and maybe take it to the places where civil society organizations are rather than assuming that they'll get it or that they'll be interested and kind of come to where the debate's already happening?

SPEAKER_02 43:10

Yeah, yeah, no, I think that's uh uh absolutely one of the one of the most important uh aspects or areas in which we should uh really bring uh AI that is capable, say, for example, uh to help uh this uh sort of democratic forums, right, where people get together and you know it could help uh finding consensus or extrapolate the most important points and see the different angles that people express and how to bring them together, right? So that would be uh a very interesting uh usage of uh natural language processing, for example. Uh but then you don't want people to be forced to give, let's uh put it uh like same quasi or semi structured interviews, right? You want them to be able to uh just brainstorm, co create, uh really throw in every any idea that comes out, and then have the I create sort of a structure.

SPEAKER_03 44:00

of all the ideas right you don't want the people to be the in the first place you know have a script and try to within this narrow script uh find some something interesting because then you would actually limit the uh the the um uh actually the the outcomes you can get right the quality of the outcome i'm not sure if uh this example speaks to the uh to to to the to your question roy but i think that's one of the ideas that uh that came to mind yeah on my side uh sometimes and sometimes i get slightly angry with all my all my ethicist friends who who uh who keep uh sounding the alarm bell but um but we we do we do need people like that i mean we need people who are thinking ahead and who are trying to portray uh the worst case scenario if we do nothing it's I think that's um that's uh yeah it's super important and sometimes uh if people and like you said civil society organizations need the visceral motivation to act well so be it is just part of uh part of our uh toolkit I guess yeah absolutely and one of the things I want to ask maybe it's sort of related to that I think a challenge that a lot of organizations that aren't you know operating in the world of AI or or particularly kind of within um spheres of the of digital civil society find when it comes to understanding AI is they feel as though they need to make a judgment about how how much store to place in some of the claims that are made about AI. Do you think you know we've talked about the fact there there clearly are really powerful potential use cases, but there is also undoubtedly a lot of hype in the wider world of of AI. How much of a challenge do you think it is for philanthropic organisations to navigate that and to understand what is real and meaningful and that they should engage with and what they should maybe you know sit back a bit and and wait and see because it feels as though there's so much heat and noise around this that it's often very difficult to make those those judgment calls.

SPEAKER_01 45:54

To me the the fundamental question is one of attitude and the the best attitude that I would recommend is uh the one that sees ai as simply a tool. So it's just a tool. It's a tool and then you judge whether it's it's useful for you whether whether you can deploy it in your organization. Like every tool it comes with uh with pluses and minuses and uh and I think that's the just that's simply the best way to anchor anchor the discussion. So that immediately kills all the hype if AI is not useful for you uh right now you just don't use it uh just keeping keeping the conversation prosaic and practical and if I can uh just plug one thing uh we uh the Altress league created a uh a very sort of common sense toolkit for deploying ai in um in philanthropic organizations in foundations and nonprofits and it's completely completely free to use and then you can download it on our on our website or just go to readyforai.org and and it's uh it's uh yeah the the the the point of view is very sort of pragmatic and and it uh um we believe that uh that the uh the the initial discussion and initial sort of sort of uh understanding of what ai can do for the organization the initial uh kind of uh education of the of the staff the initial communication consensus building can take place in a week and that's and that first phase the most important phase um once it's done then you are kind of you're already already set on your on our ai use uh trajectory so i wouldn't uh yeah I wouldn't I wouldn't overcomplicate uh a field that is already quite complex for the for different reasons when it comes to use you have an organization you have organizational goals and um if an instrument serves those goals great use it if it doesn't no need yeah I think also to to complement the the good thing of all these digital tools AI or not is that they're fairly cheap to experiment with right so you can try say uh the open AI uh LLM and if you don't like it you can try the mistral one or you can you know it's normally it's quite fairly cheap.

SPEAKER_02 48:00

The most thing you have to do is probably is just to create an account and then try uh to just try it in the right way right with a fair uh you know a cautious uh approach uh for example there have been cases of NGOs uh sharing sensitive information through these tools not realizing that the tool is actually absorbing every information every single piece of data you put in so just make the first sort of the one to three or look left and right uh approach to uh sort of watch out for the most important dangers and uh learn how to use these tools properly and then just uh experiment right but then I completely agree that we are in really in a very very deep jungle of tools that you for the one thing you have 80 different tools and you have no idea which one is better or not. And for each of these you have maybe the same number of reviews that tell you these five are the better ones or that five are the better ones. So it's really easy to feel overwhelmed and lost. But maybe initiatives like the book right not to to promote it too much but say you know that it's trying to give uh sort of some examples this is how some people uh which tools have been used is how some people do it or these are the ethical principles to keep in mind or the risks uh to be aware of such initiatives doesn't have to be this there's probably uh thousand different ones but people can use to sort of uh uh get um get the the the started with the right foot and it's a bit like taking uh uh a photographic course right nowadays you can take a million and maybe you get lost at the second time you you know turn the wheel on your camera to see this function that seems really complex but then you know if you're sort of systematic and set small goals and try to uh sort of uh have a um an educated approach to it then it's uh probably quite easy to find the tools that uh can help you and and to use them in the right way so I think the there's always this uh first move uh cost to pay that you know if it feels feels like you're climbing Mount Everest but then actually maybe it's just a small hill uh then you know if you get used to it then you'll be uh you'll be able to to to use the the AI tools so just losing sort of uh this fear and then just to add to that and again like Giuseppe said not to over overpromote the book but uh if the challenge is simply keeping abreast of of what's going on and uh listening to some smart voices in in in the jungle um take the book see all the chapter authors and uh follow those people and I think you're gonna have a fantastic overview of where AI is headed uh in terms of its relationship with uh philanthropy so that would be a piece of practical advice yeah absolutely that would be a good follow list so I think that's a good action point for anybody uh listening and I guess just to just to round off what what both of you are saying there I think it touches on something really important which is I think in these conversations sometimes we make an assumption that this is all about things happening at an organizational level and then filtering down.

SPEAKER_03 51:02

So it's about organizations and their boards of trustees deciding to adopt a particular tool and then sort of purchasing an IT solution in a relatively old fashioned way. But when it comes to AI it feels as though that's that's not actually the case most of the time these tools are out there and are often freely or cheaply available. And so individuals within organizations are probably using them already even if the organization is not thinking about it. So actually it's working from from the ground up. So I mean to to the point you were making before Miros actually we're past the point where there's a choice about whether or not to engage with this because most organizations need to just realize that you know they need to know that their their employees are going to be using tools like ChatGPT and others even if they as an organization haven't got their thinking clear on it yet. So I think that's quite a sort of important point. And then the the final question I had really was just about where things go from here. I mean one of the the great things I think about the book is that it gives a really good overview of kind of current thinking on the relationships between philanthropy and AI in all sorts of different directions. So given that what's your sense of you know what do we know that we still don't know where are there gaps in terms of our knowledge and where would you like to see more research about philanthropy and AI over the coming years?

SPEAKER_02 52:17

Giuseppe maybe if you take that first yeah so the the first thing I would like to see is more research per se from academia in developing tools and finding ways to use existing tools to assist AI uh to assist AI philanthropy through AI or through machine learning other approaches that rely on on big data crunching. I think that's uh one of the uh first points right is trying to get more people engaged uh from the from the digital world to the philanthropic sector and this is uh something that I think will happen uh so the philanthropy research per se is quite a new uh I would say academic uh focus uh there's not there's only one school of philanthropy I don't know how many business schools exist right as far as I can tell there is only the Lively School of Philanthropy as an entire school dedicated to that but we're seeing more more and more chairs or departments getting a quite a nice attention to philanthropy. So hopefully this will also mean that the hottest trends like AI will be studied in the context of philanthropy than providing more academic references for philanthropic professionals and practitioners than to to know which tools exist and how to do philanthropy in a certain property in a certain way or in another way and so on. And with respect to um the philanthropic sector as I think also thanks to this capacity building or thanks to this realization of the necessity to embrace AI, there will be also this sector bringing in their um expertise and their views on how to achieve responsible AI, how to promote responsible AI so that it's not only up to sort of market forces that are just maximizing their own interests in most cases to develop these tools. So then basically enforce regulations or try to hone better regulations for the two uh you know ad hoc situations in different parts of the of the world and so on. So that's uh the two main directions that I see that uh um should should attract attention or I hope will attract attention uh by researchers and professionals alike.

SPEAKER_01 54:26

Yeah on my side I I fully agree if I can add one thing when I think about the future of philanthropy I see it in this triangle between um trust based giving uh focus on systemic change and uh uh the use of artificial intelligence because if you want to do the first two you need the the third so I see them simply as going together and I I hope that AI will be increasingly de-hyped and just treated as a as a useful tool as we as as as the years progress. In terms of the more broadly um the trends in in the industry uh you I fully agree with what you said that uh this trend is slightly different uh because it's uh uh the the benefits of AI used accrue at the bottom not so much at the at the very top but I still I still do feel I mean what what I mean by that is they they empower the the the employees uh as opposed to the management but I still do feel that um we will see uh uh some of the aspects of some of the previous technology trends uh for example uh the trend of business process reengineering and the so-called productivity paradox in the 70s and the 80s and what we saw there was lots of investment and a certain lag in terms of get seeing those productivity figures and the reason why we we were seeing a lag was because obviously you cannot just put technology into an organization and hope for the best there needs to be some some reorganization there needs to be some education and so on. So I think um uh executives putting AI to use in philanthropic organization and elsewhere need to be ready for that. So they need to be ready to see um a certain lag in in in achieving those productivity gains. They need to be ready to communicate on what the vision is they need to be especially ready to listen to the people from the ground because they will more likely than not be the champions and those doing the the the most advanced experimentation with these uh with these new new tools and uh they need to be ready to to adapt course as this uh wild uh sector of ours uh keep keeps changing as the as the technology adapts absolutely and I think a great thought on on which to leave things and just remain to say thanks ever so much to both of you for for coming on the podcast.

SPEAKER_03 56:42

I'll obviously put links in the show notes to places where people can find the open access version of the the book and all of the the chapters contained in it and I'll put references to uh links to some of the other things that we've talked about um during the podcast. Yeah and just yeah thanks ever so much for for coming on all the best with all of your your future efforts. I'm sure we'll continue this conversation around philanthropy and and AI um you know and I'm sure everybody listening will agree it's an absolutely fascinating field with loads of really interesting work going on at the moment.

SPEAKER_02 57:10

It was a pleasure Rodri. Thank you very much for hosting us Rodri.

SPEAKER_03 57:14

Great pleasure talking to you hello it's Rodri again here. And before we get into the end matter of this podcast and I give you the usual uh sort of parish notices I thought I'd just take a couple of minutes to let you know a little bit more detail about the chapter that I've uh had published in the the Ratlich hand book on on AI and philanthropy. And so my chapter is uh called Guided Choices the Ethics of Using Algorithmic Systems to shape philanthropic decision making which tells you most of the things that you need to know about it to be honest. But essentially what I'm looking at is the question of how can AI and algorithmic tools potentially be used to shape the choices that people make when it comes to deciding to give to good causes. And I guess the reason that I'm interested in this is it feels underexplored at the moment as an aspect of the whole question of the intersection of philanthropy and AI. It's noted as I say in the chapter in other areas that there's a lot of focus on the way in which AI can be used to make decisions about us as individuals but there's less focus up to this point on the way in which AI can potentially affect the decisions we ourselves make. And the the argument that I outline in the the paper and the chapter is that you know choice is a pretty fundamental aspect of philanthropy. There's obviously a deep philosophical discussion to be had about whether philanthropy is actually a choice or a or a duty or responsibility but assuming that to some extent there is an element of choice in philanthropy because people are voluntarily choosing to give away their private assets for the public good then I think we can kind of agree that that choice is something that we need to be aware of and we need to be aware of the factors then that influence those choices because again you know there's a model within classical economics particularly of choice as something that is undertaken by purely rational actors who are trying to maximize their own utility or value in one way or another. But the reality is that actually choice is influenced by a whole range of dis different factors both kind of conscious subconscious and unconscious and that's particularly true when it comes to charitable giving and you know a lot of the the literature that explores this identifies multiple different things that that kind of play a role in informing the choices we make. So I think choice is really important when it comes to philanthropy. As I argue in the the chapter I think choice is also a really interesting concept when you start thinking about AI because there are various ways in which AI I think can affect the way that we make choices. And what I focus on in the paper is three particular ways that it can do that. So the first is the use of algorithms to determine responses to requests for information. So thinking here about you know the way in which search internet search and the algorithms that power that is already functioning because obviously you know the the information that is returned to us and the form in which it is returned to us has a big bearing then on the actions that we take as a result. And I think you know search is already determined by algorithms as I think we all know at this point. I think anybody who thinks that you know the the the results you get from a Google search are just a kind of purely objective list of the most relevant things in in answer to a particular search term is being quite naive. You know, those things are clearly determined by you know very, very powerful and lucrative algorithms that have been designed very carefully over time. And also I think increasingly there's an awareness that there's a shift going on maybe away from the traditional model of search as something that returns a list of possible results and which you can then scroll through and go down to the next page towards a model where we are being presented with answers in the form of either recommendations or things that look more like a kind of natural language conversational answers. That's already happening with online search so you will maybe have noticed in the last six months or so quite a big shift with Google towards um you know the first result that comes up with most search terms now is a kind of potted AI produced summary explaining in a kind of you know essay format or a kind of small briefing the the answer to your question. And I think a lot of people will probably rely on that. And also the more that people are using other forms of interface, conversational interfaces for instance Alexa or serial that kind of thing, the more that they are then getting you know answers in the form of kind of single responses or a small handful of responses rather than a long list. So I think you know AI is having a very clear effect there. I think beyond that I also look at the capacity of AI to enable a high degree of personalization and sort of tailoring of information and recommendations and even you know to to engage in what's known as hyper nudging so sort of to go beyond traditional nudge theory where you kind of shape choices and and kind of provide particular choice architectures to try and get people to to do specific things or undertake particular actions. And hyper nudging is essentially the the kind of AI powered souped up version of that where rather than a one-off nudge that you put in place like for instance you know where you position food in a cafeteria which is a sort of classic example in behavioral economics you can have nudges where somebody is being nudged to do something when they um you know they kind of do uh undertake an activity or search for something online but then the nudge uh is able to adapt and kind of respond dependent on what actions they take so it's potentially much much more powerful um and then the third thing that I look at is the use of AI generated content uh to prompt emotional responses and drive behavior in that way. So the question is sort of how do you use generative AI and and imagery and potentially sort of you know video and and that kind of thing in the future to try and get people to give more to charity or sort of you know do have an empathetic response. So I explore all of those in the paper and essentially without you know spoiling it what I do is kind of look through some of the potential ethical issues that these raise um in terms of you know the main things being the impact that this might have on our individual autonomy and whether that is a price worth paying, you know, particularly when it comes for instance to providing more tailored recommendations or suggestions, you know, the argument would be in the context of philanthropy perhaps that would drive more giving and maybe it would result in more effective or better distribution of giving overall but the counter is if that is done in a way that undermines people's individual agency and autonomy is that something that we should be concerned about both inherently for itself if we value autonomy and agency and potentially because there are longer term unintended consequences to that. Similarly I think in terms of the use of generative AI, you know imagery and other content to try and drive emotive responses, even if that is found to be effective what is the potential wider cost in terms of undermining people's trust in what they see and hear online and what might the ramifications of that be for civil society in particular? And also you know can we even assume that it is effective? I mean there is a bit of um research already suggesting that when people are aware that content has been generated using artificial intelligence, even if you know the people doing so are very uh clear and upfront about that, they actually are less inclined to give for instance in the context of a fundraising request. So I think there are reasons to be cautious about some of the the the kind of you know suggestions that we might go full steam ahead in in using some of these tools in those contexts. And then to finish off the chapter I sort of outline what some of these key ethical questions might be and then what that might mean in terms of the response of the non profit sphere and philanthropic organisations and policy makers more widely in terms Of uh addressing them. So if you're interested in that, I'll put links in the show notes. Do check it out. I think it's quite an interesting paper, if I do say so myself, you know, and if you don't, feel free to disagree with me, but do so in an interesting way. I'm always up for that. Well, that brings us to the proper end of the podcast. Uh, just remains to say thanks ever so much again to Milos and to uh Giuseppe for coming on the podcast. Great to have a chance to talk to both of them. Um, it's been really great to be involved in this project uh and also kind of you know involved in the group that has um kind of uh developed around uh AI and philanthropy as a result of the work that they've both been doing. Um and certainly if any of you are part of that group and were at the conference uh in Geneva earlier this year, hello to you. I'll put links uh in the show notes to places where you can find uh the book itself, which I should say is currently available open access, so you can read all of those um chapters online, and there are loads and loads of great chapters covering all sorts of different aspects of the intersection of philanthropy and AI, so do you know carve out a bit of time to do that. Um I'll put some links to other bits and pieces that I've written or said about AI that you might be interested in as well. If you're interested more broadly in philanthropy and civil society, and you like the kind of thing that I bang on about, then you know do take some time to have a look through the website at why philanthropymatters.com. You can find all the back episodes of this podcast day, you can find lots of articles and news updates and short guides to all kinds of things to do with um philanthropy, and you know, there's many of them that touch on issues around technology. Uh, I even uh quite recently added a new guide on uh the question of philanthropy and quantum technology. So if you're interested in some highly speculative thoughts on on an emerging technology and its relevance for philanthropy, do check that one out. If you're interested in more you know up-to-date uh kind of bits of news from me, do follow me on social media or uh kind of uh connect with me. I'm on LinkedIn at the moment. I'm also trying to use Blue Sky a bit more, so do come and find me on there and try and make that more of a party so that we can uh all kind of um finally totally bin off X or Twitter or whatever nonsense we're calling it these days. Other than that, it just remains to say like, subscribe, follow the podcast, do leave us a nice review wherever it is that you get your podcasts. Maybe tell somebody that you think might be interested in it, because I think personal recommendations go a long way. Uh, and other than that, I'll see you next time. Bye.