WEBVTT 00:00:00.080 --> 00:00:03.660 Hello and welcome to the very first episode of 00:00:03.660 --> 00:00:08.519 our new podcast series, The Purpose Brief. It's 00:00:08.519 --> 00:00:10.759 a space to help you connect with everything to 00:00:10.759 --> 00:00:13.820 do with creating more social good in the world. 00:00:13.919 --> 00:00:20.109 What's not to like? To kick things off, I'm joined 00:00:20.109 --> 00:00:23.370 by Nathan Chappell, a trailblazer at the intersection 00:00:23.370 --> 00:00:26.829 of philanthropy and technology. He's chief AI 00:00:26.829 --> 00:00:30.190 officer at Virtuous, founder of Fundraising AI, 00:00:30.570 --> 00:00:33.670 and co -author of The Generosity Crisis. It's 00:00:33.670 --> 00:00:36.329 a book that says the real challenge for modern 00:00:36.329 --> 00:00:39.409 giving isn't really about the lack of money out 00:00:39.409 --> 00:00:41.609 there in the world. It's a lack of connection, 00:00:41.929 --> 00:00:44.840 human connection. Nathan's also just released 00:00:44.840 --> 00:00:46.600 a new book, which we'll talk a little bit about, 00:00:46.759 --> 00:00:50.159 called Nonprofit AI, which explores how artificial 00:00:50.159 --> 00:00:53.179 intelligence can be used responsibly to strengthen 00:00:53.179 --> 00:00:56.679 trust, empathy and impact across the social sector. 00:00:57.420 --> 00:01:00.920 He's spent over two decades exploring how technology 00:01:00.920 --> 00:01:04.900 can help us give better, not just faster, and 00:01:04.900 --> 00:01:08.260 how to keep generosity human in a world run by 00:01:08.260 --> 00:01:11.659 those pesky algorithms. He's an inventor, he's 00:01:11.659 --> 00:01:14.680 a dad, and someone who still finds time for a 00:01:14.680 --> 00:01:17.340 long cycle ride to clear his head at the end 00:01:17.340 --> 00:01:20.379 of the day without his phone. It's a thoughtful 00:01:20.379 --> 00:01:23.120 conversation to start the series about trust, 00:01:23.319 --> 00:01:26.379 about belonging, and what generosity might look 00:01:26.379 --> 00:01:29.819 like in the years ahead. So you ready to join 00:01:29.819 --> 00:01:33.060 us for this journey? Let's dive into the first 00:01:33.060 --> 00:01:39.159 episode. Nathan, hello. How are you? I'm doing 00:01:39.159 --> 00:01:41.140 well, James. How are you? Really good to see 00:01:41.140 --> 00:01:44.120 you. Thank you so much for joining us on what 00:01:44.120 --> 00:01:46.680 is essentially the first edition of our Purpose 00:01:46.680 --> 00:01:49.859 Brief podcast. Thank you very much for agreeing 00:01:49.859 --> 00:01:52.219 to do it. Yeah, absolutely an honor. It was great 00:01:52.219 --> 00:01:55.180 to be able to meet you up in York. And, you know, 00:01:55.200 --> 00:01:57.019 I'm so glad you reached out. Congratulations 00:01:57.019 --> 00:01:59.420 on your book. And that's been out for a few months 00:01:59.420 --> 00:02:03.500 now. Let's talk about our first top level question. 00:02:03.959 --> 00:02:07.060 What is your purpose? What do you do? That's 00:02:07.060 --> 00:02:08.819 a big question to start with. Like, what is your 00:02:08.819 --> 00:02:10.800 purpose? I don't know if you've ever seen the 00:02:10.800 --> 00:02:13.139 movie The City Slickers. It was an old movie 00:02:13.139 --> 00:02:15.520 I grew up in. Well, I grew up in, I was born 00:02:15.520 --> 00:02:17.120 in the seventies, but probably grew up in the 00:02:17.120 --> 00:02:19.460 eighties movies. And there was this question 00:02:19.460 --> 00:02:22.219 in, you know, it was like, what is your one thing, 00:02:22.219 --> 00:02:26.039 your one thing in life? And that really, like, 00:02:26.120 --> 00:02:27.780 I don't know. I grew up at a time where that 00:02:27.780 --> 00:02:30.180 movie was formative and I, it bothered me that 00:02:30.180 --> 00:02:32.479 I didn't know my one thing. And so if people 00:02:32.479 --> 00:02:34.240 don't know this movie, it essentially is Billy 00:02:34.240 --> 00:02:38.139 Crystal. And he goes on this little escapade, 00:02:38.139 --> 00:02:41.439 a ranch, and he essentially finds his one thing 00:02:41.439 --> 00:02:44.000 in life. For the most part, I was drawn to philanthropy 00:02:44.000 --> 00:02:46.479 early on at a very young age. Actually, I did 00:02:46.479 --> 00:02:48.159 my first fundraiser when I was eight years old, 00:02:48.219 --> 00:02:50.599 but I never thought about philanthropy as a career. 00:02:50.819 --> 00:02:53.419 I landed into it as an accidental fundraiser 00:02:53.419 --> 00:02:55.900 in the year 2000, ended up raising money for 00:02:55.900 --> 00:02:58.620 many years. And it took me a while to understand 00:02:58.620 --> 00:03:02.099 my one thing. And I think it actually came to 00:03:02.099 --> 00:03:07.000 me when I was provided an opportunity to leave 00:03:07.000 --> 00:03:09.860 my job. And when I left, I remember telling the 00:03:09.860 --> 00:03:13.289 person that, I worked for that I was really driven 00:03:13.289 --> 00:03:16.710 by the idea of scaling generosity. At the truest 00:03:16.710 --> 00:03:20.569 sense, my purpose was to scale generosity and 00:03:20.569 --> 00:03:22.889 being a technologist, someone who had started 00:03:22.889 --> 00:03:24.810 several tech companies, sold those companies, 00:03:24.949 --> 00:03:28.430 super early adopter, always curious about what's 00:03:28.430 --> 00:03:32.189 next. It was really this idea of scaling generosity 00:03:32.189 --> 00:03:35.389 through technology. And so that's been really 00:03:35.389 --> 00:03:39.289 my purpose ever since I came across that. It 00:03:39.289 --> 00:03:44.210 just felt my mantra in life. And while I believe 00:03:44.210 --> 00:03:46.229 humans are amazing and they're capable of amazing 00:03:46.229 --> 00:03:48.210 things, they're not scalable on their own. So 00:03:48.210 --> 00:03:50.289 it's really this idea of the intersection of 00:03:50.289 --> 00:03:54.069 generosity and AI, which I've been on this path 00:03:54.069 --> 00:03:56.889 since about 2017 now. A great movie reference 00:03:56.889 --> 00:04:00.370 there. When you talk about social good, what 00:04:00.370 --> 00:04:02.349 does it mean to you? Because we're obviously 00:04:02.349 --> 00:04:05.789 very passionate about through our business, but 00:04:05.789 --> 00:04:08.930 what does it mean to you specifically? Yeah, 00:04:08.949 --> 00:04:10.889 I mean, social good is a broad term that can 00:04:10.889 --> 00:04:15.129 mean anything. I mean, ultimately, what I was 00:04:15.129 --> 00:04:18.889 raised to believe and what I, you know, believe 00:04:18.889 --> 00:04:21.970 as an adult and instilled in my children is that, 00:04:21.970 --> 00:04:24.029 you know, no matter what you have, there's always 00:04:24.029 --> 00:04:27.149 a responsibility to, you know, there's always 00:04:27.149 --> 00:04:29.089 opportunity and responsibility to give back. 00:04:29.189 --> 00:04:31.730 And so, you know, social good for me is really 00:04:31.730 --> 00:04:34.170 a state of mind, not something that you do. It's 00:04:34.170 --> 00:04:37.189 a way of being. And, you know, when I became 00:04:37.189 --> 00:04:40.410 an accidental fundraiser, in fact, I fell into 00:04:40.410 --> 00:04:43.569 this on accident. But for my first seven years 00:04:43.569 --> 00:04:46.290 running a really small grassroots nonprofit, 00:04:46.930 --> 00:04:49.329 I never felt like I went to work. I felt like 00:04:49.329 --> 00:04:52.389 it was a way of life. And, you know, from a biological 00:04:52.389 --> 00:04:56.110 perspective, humans are these, you know, pro 00:04:56.110 --> 00:04:59.850 -social creatures that tend to like biologically 00:04:59.850 --> 00:05:02.870 we're wired to actually help others even at sometimes 00:05:02.870 --> 00:05:06.290 a cost to us. So for me, social good is tying 00:05:06.290 --> 00:05:09.509 into that primal instinct. It's something that 00:05:09.509 --> 00:05:13.310 we as humans are intended to do, to give back. 00:05:13.329 --> 00:05:16.069 This is why our brain releases dopamine and serotonin 00:05:16.069 --> 00:05:20.050 when you help. It's why we release oxytocin when 00:05:20.050 --> 00:05:23.569 we have empathy for others. I mean, we're biologically 00:05:23.569 --> 00:05:27.370 built for the idea of social good. It's really 00:05:27.370 --> 00:05:31.139 the noise and the distraction in societies. that 00:05:31.139 --> 00:05:34.060 keep us from living in that kind of truest sense. 00:05:34.199 --> 00:05:36.819 But, you know, I think that is probably the foundational 00:05:36.819 --> 00:05:39.579 belief are people, you know, inherently good 00:05:39.579 --> 00:05:42.600 or evil. I don't know. I mean, I guess you'd 00:05:42.600 --> 00:05:45.040 go back to the biblical days and, you know, Adam 00:05:45.040 --> 00:05:48.839 took a, you know, bite of the apple. But, you 00:05:48.839 --> 00:05:51.759 know, at the end of the day, I do believe that 00:05:51.759 --> 00:05:54.680 people are wired to give back and they're wired 00:05:54.680 --> 00:05:57.079 to be in community. And so social good for me 00:05:57.079 --> 00:06:00.310 is the expression of that, that goodness. So 00:06:00.310 --> 00:06:03.970 in a way, your work in AI is a kind of counterpoint 00:06:03.970 --> 00:06:07.850 to that in the sense that in the world of impacted 00:06:07.850 --> 00:06:11.350 by algorithms that we're starting to see that 00:06:11.350 --> 00:06:14.850 has a negative effect on community and kind of 00:06:14.850 --> 00:06:18.629 cohesion. You feel that through your work, AI 00:06:18.629 --> 00:06:22.610 can be a force for good. Is that against algorithms 00:06:22.610 --> 00:06:25.350 and everything that social media is unlocking? 00:06:26.120 --> 00:06:27.920 Yeah, that's a great question, you know, and 00:06:27.920 --> 00:06:31.920 I get that a lot in the sense of that belief 00:06:31.920 --> 00:06:34.100 that maybe, you know, technology and automation 00:06:34.100 --> 00:06:38.439 is, you know, morally in opposition of community. 00:06:39.040 --> 00:06:42.300 And maybe it's because I like a challenge and 00:06:42.300 --> 00:06:45.560 I, you know, don't like to do easy things. But 00:06:45.560 --> 00:06:48.740 to be honest, like I think about AI as any other 00:06:48.740 --> 00:06:51.240 technology, like, I mean, no different than, 00:06:51.240 --> 00:06:55.750 you know, a hammer or a saw or. typewriter. It's 00:06:55.750 --> 00:06:57.689 a technology that doesn't know the difference 00:06:57.689 --> 00:07:00.329 between right or wrong. It's essentially in the 00:07:00.329 --> 00:07:02.629 hands of humans that will either use it for good 00:07:02.629 --> 00:07:07.009 or bad purposes. And so while inevitably there 00:07:07.009 --> 00:07:10.149 could be more bad actors than good actors in 00:07:10.149 --> 00:07:12.990 the world, or those bad actors will use this 00:07:12.990 --> 00:07:15.529 technology in ways that will not be beneficial 00:07:15.529 --> 00:07:18.949 to humanity, I do believe that the technology, 00:07:19.050 --> 00:07:22.170 again, that doesn't know the difference. can 00:07:22.170 --> 00:07:25.110 be a force multiplier for, especially for social 00:07:25.110 --> 00:07:27.550 good, for, you know, social impact organizations. 00:07:28.350 --> 00:07:30.649 And when you think about it, like humans are 00:07:30.649 --> 00:07:34.430 amazing. Like we are, I host dinners when I travel 00:07:34.430 --> 00:07:36.550 called dinners of extraordinary humans, because 00:07:36.550 --> 00:07:38.930 I believe that every human, like we're, we're 00:07:38.930 --> 00:07:41.670 walking miracles. Like we are like, there's no 00:07:41.670 --> 00:07:44.610 reason why like we're even here and having consciousness 00:07:44.610 --> 00:07:47.069 and all these like amazing thoughts and ideas. 00:07:47.129 --> 00:07:49.420 And we can do these things. We are extraordinary, 00:07:49.500 --> 00:07:52.699 but we're not scalable on our own. We have always 00:07:52.699 --> 00:07:55.339 needed technology. Humans have always, progress 00:07:55.339 --> 00:07:59.120 is always capped by the technology that lets 00:07:59.120 --> 00:08:01.420 us amplify our intentions. And so whether it 00:08:01.420 --> 00:08:04.899 be electricity or whether it be the typewriter 00:08:04.899 --> 00:08:07.319 or the printing press or whatever it might be, 00:08:08.259 --> 00:08:11.620 while humans are incredible, we need technology 00:08:11.620 --> 00:08:15.800 to amplify. And so I think AI is the greatest 00:08:15.800 --> 00:08:18.939 force multiplier for... the amplification of 00:08:18.939 --> 00:08:22.980 ideas if used well. And so while I'm very bullish 00:08:22.980 --> 00:08:26.259 on the idea of AI being used for good, I'm also 00:08:26.259 --> 00:08:29.779 deeply concerned that with a lack of information 00:08:29.779 --> 00:08:33.799 or without a lot of foresight or thought about 00:08:33.799 --> 00:08:36.860 unintended consequences, that the technology 00:08:36.860 --> 00:08:40.879 could even purposefully or accidentally be used 00:08:40.879 --> 00:08:45.399 in a way that moves people apart. And so I find 00:08:45.399 --> 00:08:48.809 it... Most of my passion work, my full -time 00:08:48.809 --> 00:08:53.009 volunteer job now with fundraising .ai is educating 00:08:53.009 --> 00:08:56.169 people, not advocating for the use of AI, but 00:08:56.169 --> 00:08:58.289 advocating for the responsible and beneficial 00:08:58.289 --> 00:09:01.250 use of AI, which means that it supports us both 00:09:01.250 --> 00:09:03.629 in the short term and in the long term. You've 00:09:03.629 --> 00:09:07.730 just completed the summit that you had over the... 00:09:07.929 --> 00:09:09.870 I think it was about a month ago, wasn't it? 00:09:10.009 --> 00:09:12.809 And that was kind of record numbers for you. 00:09:13.250 --> 00:09:15.450 People from all across the world from not -for 00:09:15.450 --> 00:09:18.169 -profits getting together. Was there a sort of 00:09:18.169 --> 00:09:21.190 takeaway in terms of the outcomes and the discussions? 00:09:21.909 --> 00:09:24.389 You know, it's interesting because fundraising 00:09:24.389 --> 00:09:26.710 AI started out as this, you know, grassroots 00:09:26.710 --> 00:09:30.190 volunteer effort. Well, really 2018 and no one 00:09:30.190 --> 00:09:32.490 really cared about it, to be honest. Like AI 00:09:32.490 --> 00:09:34.750 was not very accessible. It's not very affordable. 00:09:35.659 --> 00:09:38.320 And so, I mean, we're talking, we're a few hundred 00:09:38.320 --> 00:09:41.960 people. And then in 2022, Chanchi PT came out. 00:09:42.039 --> 00:09:45.340 I gathered about, I don't know, 75 people. We 00:09:45.340 --> 00:09:48.320 decided to do this summit. And a few months later, 00:09:48.419 --> 00:09:50.960 we had our first summit, which brought, we thought 00:09:50.960 --> 00:09:53.460 about 500 people would come and we had 5 ,000 00:09:53.460 --> 00:09:55.639 people. And then the next year, 7 ,500 people. 00:09:55.759 --> 00:09:58.179 This year, 11 ,000 people from 130 countries. 00:09:58.879 --> 00:10:03.419 I've seen the evolution from. total skepticism 00:10:03.419 --> 00:10:07.139 and people probably hopefully hoping that AI 00:10:07.139 --> 00:10:10.059 was never going to come. It would be a fad to 00:10:10.059 --> 00:10:14.519 curiosity. And now we see, you know, a kind of 00:10:14.519 --> 00:10:17.080 widespread adoption, but not in any strategic 00:10:17.080 --> 00:10:19.299 way. And so I think the observations from this 00:10:19.299 --> 00:10:22.740 year are we've crossed the barriers of for the 00:10:22.740 --> 00:10:26.230 people who resist change, which are. 30 % of 00:10:26.230 --> 00:10:28.909 people resist change. 50 % of the people sit 00:10:28.909 --> 00:10:30.730 on the sidelines. So, you know, we're talking, 00:10:30.850 --> 00:10:33.970 you know, a vast majority, 80 % of people are 00:10:33.970 --> 00:10:36.570 not going to be your, your, you know, front runners 00:10:36.570 --> 00:10:39.269 in this. They're now brought along. We're in 00:10:39.269 --> 00:10:42.190 this like middle of the adoption, but there's 00:10:42.190 --> 00:10:43.850 not a lot of strategic use. They've accepted 00:10:43.850 --> 00:10:47.490 that this isn't going away, but it's still very 00:10:47.490 --> 00:10:51.330 much in its kind of infancy in terms of using 00:10:51.330 --> 00:10:54.600 it in ways that are. not really formative. There 00:10:54.600 --> 00:10:57.500 are ways that are just slightly, you know, incremental 00:10:57.500 --> 00:11:01.240 on the sides. And to be honest, with the headwinds 00:11:01.240 --> 00:11:03.259 facing our sector, societal changes facing our 00:11:03.259 --> 00:11:06.120 sector, incremental changes don't work in an 00:11:06.120 --> 00:11:08.919 exponential world. And so that's the part where 00:11:08.919 --> 00:11:11.519 I think the fundraising .ai kind of movement 00:11:11.519 --> 00:11:13.659 is really to help educate people to get people 00:11:13.659 --> 00:11:17.659 away from just using shadow AI to improve a letter 00:11:17.659 --> 00:11:20.200 once in a while to changing their orientation. 00:11:21.210 --> 00:11:24.309 to thinking from the first bit of work they do 00:11:24.309 --> 00:11:27.129 in the morning is how can AI help? And do you 00:11:27.129 --> 00:11:31.450 think that adoption changes the human purpose 00:11:31.450 --> 00:11:33.789 within the organizations? I mean, I think that 00:11:33.789 --> 00:11:36.490 worry that people will have or a lot of people 00:11:36.490 --> 00:11:40.190 do around adoption and their future jobs and 00:11:40.190 --> 00:11:42.289 what does the future of work look like? I mean, 00:11:42.330 --> 00:11:46.929 is that something you see as being adopted more 00:11:46.929 --> 00:11:49.509 readily by fundraisers who already have constrained 00:11:49.509 --> 00:11:54.190 budgets? Yeah, you know, it does vary by country 00:11:54.190 --> 00:11:56.409 quite a bit because the ethos of the country 00:11:56.409 --> 00:12:01.129 and kind of the sediment around what AI represents 00:12:01.129 --> 00:12:04.750 does kind of dictate this. And so, of course, 00:12:04.809 --> 00:12:08.049 the U .S. has taken this very much rogue, like 00:12:08.049 --> 00:12:11.350 anything goes, just try everything. And when 00:12:11.350 --> 00:12:13.110 something really bad happens, we're going to 00:12:13.110 --> 00:12:16.470 regulate it. Whereas the EU is taking a much 00:12:16.470 --> 00:12:18.610 more pragmatic approach to say, well, let's. 00:12:19.120 --> 00:12:21.960 put some guardrails around AI, have people think 00:12:21.960 --> 00:12:25.039 about some of the really unintended or the bad 00:12:25.039 --> 00:12:27.700 use cases of it and not let people do those. 00:12:27.759 --> 00:12:30.679 And let's regulate before something bad happens. 00:12:30.980 --> 00:12:34.120 So it does change a little bit of that. But I 00:12:34.120 --> 00:12:37.320 do think the fear factor is starting to creep 00:12:37.320 --> 00:12:39.460 in. I mean, we're starting to see, you know, 00:12:39.460 --> 00:12:41.539 for early days, it was like, well, AI is not 00:12:41.539 --> 00:12:44.299 going to replace jobs. But people who use AI 00:12:44.299 --> 00:12:46.440 will replace those that don't. That was essentially 00:12:46.440 --> 00:12:50.039 and still probably largely the most popular belief 00:12:50.039 --> 00:12:52.620 that AI won't replace humans, but humans that 00:12:52.620 --> 00:12:55.679 use AI will replace those that don't. We're actually 00:12:55.679 --> 00:12:57.519 starting to see that change. Like we're actually 00:12:57.519 --> 00:13:02.100 seeing AI replace jobs and we're going to need 00:13:02.100 --> 00:13:04.200 to figure out like what do those people do? And 00:13:04.200 --> 00:13:06.940 now this is not new. I mean, every technological 00:13:06.940 --> 00:13:09.700 revolution has had these same types of things. 00:13:10.330 --> 00:13:13.289 But never at this scale and never at this pace. 00:13:13.490 --> 00:13:16.509 And so I do think that there is quite a bit of 00:13:16.509 --> 00:13:20.289 fear, but the bar is being raised at the same 00:13:20.289 --> 00:13:24.190 time. So fundraisers that are using AI to elevate 00:13:24.190 --> 00:13:28.110 their ability, whether it's, you know, practicing 00:13:28.110 --> 00:13:31.090 or essentially rehearsing an ask before you go, 00:13:31.269 --> 00:13:34.389 whether it's, you know, really dialing in, understanding 00:13:34.389 --> 00:13:37.309 donor sentiment and dialing in communication 00:13:37.309 --> 00:13:39.980 to a donor, whatever that might be. I do think 00:13:39.980 --> 00:13:43.360 the bar is starting to increase. And I think 00:13:43.360 --> 00:13:47.320 those that are really diving into this will very 00:13:47.320 --> 00:13:49.559 quickly outperform others. And so they'll find 00:13:49.559 --> 00:13:52.820 themselves at the forefront of the wave. It's 00:13:52.820 --> 00:13:55.259 really interesting talking to different types 00:13:55.259 --> 00:13:58.399 of not -for -profits and fundraising teams about 00:13:58.399 --> 00:14:00.860 how they're adopting it. And I wonder whether 00:14:00.860 --> 00:14:03.179 there's more of a challenge for big organizations 00:14:03.179 --> 00:14:06.480 who will be much... bigger in terms of the governance 00:14:06.480 --> 00:14:09.679 expectations versus small not -for -profits. 00:14:09.779 --> 00:14:11.480 I mean, you talked about a few great examples 00:14:11.480 --> 00:14:14.799 of somebody who is the chief exec of a charity 00:14:14.799 --> 00:14:18.600 in the US and has one fundraiser and that's it, 00:14:18.700 --> 00:14:21.480 you know, to do everything. So how they scale 00:14:21.480 --> 00:14:26.440 and change their approach is much, this could 00:14:26.440 --> 00:14:29.159 make a huge difference to them in their ability 00:14:29.159 --> 00:14:32.909 to just deliver funds. versus big organizations 00:14:32.909 --> 00:14:37.370 like higher ed institutions that might be more 00:14:37.370 --> 00:14:40.649 cautious for obvious reasons, that have multiple 00:14:40.649 --> 00:14:44.330 different priorities. That's a big change. So 00:14:44.330 --> 00:14:48.210 I started working in AI in 2017 in machine learning, 00:14:48.389 --> 00:14:50.509 so predictive AI, machine learning and deep learning 00:14:50.509 --> 00:14:52.769 to predict which donors are likely to make a 00:14:52.769 --> 00:14:56.879 gift. And, you know, to be honest, only... the 00:14:56.879 --> 00:14:59.500 big organizations could afford to do that type 00:14:59.500 --> 00:15:03.779 of work. The barriers to leveraging machine learning 00:15:03.779 --> 00:15:05.960 and deep learning, either you had to hire a PhD, 00:15:06.179 --> 00:15:09.419 in early days you had to hire a PhD pretty much, 00:15:09.500 --> 00:15:13.620 or a statistician, mathematician, you know, kind 00:15:13.620 --> 00:15:17.899 of cross. And it was not easy. I mean, we spent 00:15:17.899 --> 00:15:19.580 a year and a half building our first algorithm, 00:15:19.840 --> 00:15:23.559 you know, very long -handed, lots of code. What's 00:15:23.559 --> 00:15:26.100 interesting is that... Early adopters tended 00:15:26.100 --> 00:15:28.559 to be the ones that had kind of philanthropic 00:15:28.559 --> 00:15:30.720 research and development budgets, so like philanthropic 00:15:30.720 --> 00:15:34.659 R &D. And those almost entirely, almost 100 % 00:15:34.659 --> 00:15:36.700 of those organizations were the large organizations 00:15:36.700 --> 00:15:38.539 because they were big enough that they could 00:15:38.539 --> 00:15:41.179 afford to experiment a little bit if it didn't 00:15:41.179 --> 00:15:43.120 work. And there was a lot of skepticism whether 00:15:43.120 --> 00:15:46.000 it would work or not. But that's almost reversed 00:15:46.000 --> 00:15:48.899 entirely. Like, I think that's what the huge 00:15:48.899 --> 00:15:51.600 shift has been with the advent of generative 00:15:51.600 --> 00:15:54.460 AI, which is... Basically free or nearly free, 00:15:54.519 --> 00:15:57.279 depending on what models you're using. It's put 00:15:57.279 --> 00:16:00.559 AI in the hands of everyone. And when you think 00:16:00.559 --> 00:16:03.639 about nonprofits and the idea of like need is 00:16:03.639 --> 00:16:06.360 the mother of invention, like small nonprofits 00:16:06.360 --> 00:16:08.759 with one person, two people, three people, they 00:16:08.759 --> 00:16:12.399 need to extend their reach. And AI is peanut 00:16:12.399 --> 00:16:15.059 butter and jelly for a nonprofit that is resource 00:16:15.059 --> 00:16:18.340 constrained. And interestingly enough, is that 00:16:18.340 --> 00:16:20.500 the big organizations that were first are now 00:16:20.500 --> 00:16:23.769 last. Because now all of a sudden those big organizations 00:16:23.769 --> 00:16:27.669 are mired in bureaucracy and policy. They can't 00:16:27.669 --> 00:16:31.610 agree on what to do next and what tools are approved. 00:16:31.850 --> 00:16:34.509 In fact, I have a friend who was one of the earliest 00:16:34.509 --> 00:16:37.429 adopters in AI at a fairly large children's hospital 00:16:37.429 --> 00:16:42.529 in 2017, 2018. He's using less AI now in his 00:16:42.529 --> 00:16:45.549 organization than he was four years ago. So it's 00:16:45.549 --> 00:16:47.690 reverted backward, whereas these small nonprofits 00:16:47.690 --> 00:16:49.850 are just power boosting through this movement, 00:16:49.950 --> 00:16:52.480 which is... Great to see, honestly, from a scale 00:16:52.480 --> 00:16:56.379 perspective. And do you think that's unlocking 00:16:56.379 --> 00:16:58.860 more funds for them? Is it more different types 00:16:58.860 --> 00:17:02.220 of donors? Is there any kind of key trends coming 00:17:02.220 --> 00:17:04.740 out of that? You know, I think it's starting 00:17:04.740 --> 00:17:07.660 at the lowest denominator, which is just like 00:17:07.660 --> 00:17:10.119 improving efficiency. And so for the most part, 00:17:10.759 --> 00:17:12.960 you know, it's helping those small nonprofits 00:17:12.960 --> 00:17:16.759 make the best use of their resources. And so 00:17:16.759 --> 00:17:21.470 the average. nonprofit could save 20 % of their 00:17:21.470 --> 00:17:24.630 time that provides, you know, per employee an 00:17:24.630 --> 00:17:27.289 extra day a week. And so, you know, if you could 00:17:27.289 --> 00:17:31.589 use AI to really create efficiency that gives 00:17:31.589 --> 00:17:34.609 you back 20 % of your time. And I can tell you 00:17:34.609 --> 00:17:37.589 like my second book, nonprofit AI was written 00:17:37.589 --> 00:17:39.869 in four months because I had a research assistant. 00:17:40.009 --> 00:17:43.470 I had something to bounce off ideas. I had something 00:17:43.470 --> 00:17:47.609 to help me create analogies to do. help me find 00:17:47.609 --> 00:17:50.009 case studies, my first book took two and a half 00:17:50.009 --> 00:17:53.349 years. So a nonprofit that's going to dive in 00:17:53.349 --> 00:17:54.829 and be like, hey, this is part of what we're 00:17:54.829 --> 00:17:58.009 going to do, 20 % of their time, that's a low 00:17:58.009 --> 00:18:01.289 bar. When we think about from a resource development 00:18:01.289 --> 00:18:05.509 perspective, let's say still predictive AI is 00:18:05.509 --> 00:18:09.569 the most common way of actually using AI to predict 00:18:09.569 --> 00:18:11.390 who's going to make a donation and then have 00:18:11.390 --> 00:18:13.990 more targeted outreach, but generative AI being 00:18:13.990 --> 00:18:17.380 used. for sediment analysis to really understand 00:18:17.380 --> 00:18:20.420 donor populations. And then ideally take the 00:18:20.420 --> 00:18:23.319 predictions from machine learning and deep learning 00:18:23.319 --> 00:18:27.460 and then personalize asks. And so that is still 00:18:27.460 --> 00:18:29.339 probably the most common use case right now. 00:18:29.420 --> 00:18:32.400 Second to that, I think we're seeing a lot in 00:18:32.400 --> 00:18:36.160 prospect research because I used to have four 00:18:36.160 --> 00:18:38.720 prospect researchers, full -time people with 00:18:38.720 --> 00:18:42.750 master's degrees working for me. there's zero 00:18:42.750 --> 00:18:45.490 doubt that one person now could do the work of 00:18:45.490 --> 00:18:49.150 four. And so, I mean, that is a significant power 00:18:49.150 --> 00:18:51.690 boost for nonprofits. And so even an average 00:18:51.690 --> 00:18:54.029 fundraiser who wants to be prepared before they 00:18:54.029 --> 00:18:58.230 go for an ask has this research tool that if 00:18:58.230 --> 00:19:01.130 they know how to use it well, would be the equivalent 00:19:01.130 --> 00:19:03.150 of a person with a master's degree that's going 00:19:03.150 --> 00:19:05.829 to spend several hours doing work for you. So 00:19:05.829 --> 00:19:10.069 can we imagine a future where there will be fundraising 00:19:10.069 --> 00:19:15.569 bots that a major gift manager would have working 00:19:15.569 --> 00:19:18.289 for them producing a data looking at the prospect 00:19:18.289 --> 00:19:21.309 analysis most up to date and this is what you're 00:19:21.309 --> 00:19:24.930 doing this week yeah you basically just pulled 00:19:24.930 --> 00:19:28.450 a page out of my notebook that i keep i i have 00:19:28.450 --> 00:19:31.529 a room in my house that there's no phones allowed 00:19:31.529 --> 00:19:34.769 it's it's it's a no phone zone and it's like 00:19:34.769 --> 00:19:36.789 a little library nook where i go for my first 00:19:36.789 --> 00:19:39.529 cup of coffee And especially in this day and 00:19:39.529 --> 00:19:41.670 age, I encourage everyone to do it. It's amazing. 00:19:41.829 --> 00:19:44.069 You know, if you can put that down and leave 00:19:44.069 --> 00:19:47.269 it in another room, no distraction. Yeah. About 00:19:47.269 --> 00:19:51.170 a few months ago, I was thinking about the fundraising 00:19:51.170 --> 00:19:54.390 shops that I had led. I mean, I started out in 00:19:54.390 --> 00:19:56.769 small, very small grassroots, and then I ended 00:19:56.769 --> 00:20:00.069 up my career with 250 employees, which included 00:20:00.069 --> 00:20:03.130 all fundraisers and prospect research and donor 00:20:03.130 --> 00:20:06.079 relations and all this stuff. And throughout 00:20:06.079 --> 00:20:08.880 all of those, I actually took my notebook one 00:20:08.880 --> 00:20:11.539 day and I was like, you know, I'm curious whether 00:20:11.539 --> 00:20:14.099 I had two employees where I wore lots of different 00:20:14.099 --> 00:20:17.480 hats or I had 200 employees where I wore a fraction 00:20:17.480 --> 00:20:20.740 of a hat. How many actual job functions are there 00:20:20.740 --> 00:20:23.980 in, you know, truly job functions? And so I really 00:20:23.980 --> 00:20:26.920 started like racking my brain and I. I came up 00:20:26.920 --> 00:20:29.400 with a 20 job function. So like no matter how 00:20:29.400 --> 00:20:31.779 big or how small your organization is, there's 00:20:31.779 --> 00:20:34.480 about 20 core job functions that a nonprofit 00:20:34.480 --> 00:20:38.579 person does. And then I, you know, for the purposes 00:20:38.579 --> 00:20:42.099 of the exercise was curious about in what near 00:20:42.099 --> 00:20:46.099 future state could bots essentially augment or 00:20:46.099 --> 00:20:49.220 highly, highly augment or even automate these 00:20:49.220 --> 00:20:52.900 jobs? Like how many of these 20 jobs? could be 00:20:52.900 --> 00:20:55.420 highly supported by AI. And I started with like 00:20:55.420 --> 00:20:58.539 five easy, easy prospect research, prospect, 00:20:58.660 --> 00:21:00.720 you know, donor relations, stewardship and things 00:21:00.720 --> 00:21:04.460 like that. And then I got to 11 and then I kind 00:21:04.460 --> 00:21:06.759 of got stuck at like 13. And then I came back 00:21:06.759 --> 00:21:08.779 the next day or the day after. I'm like, wait 00:21:08.779 --> 00:21:10.740 a minute. You know, literally the conclusion 00:21:10.740 --> 00:21:14.319 is that all 20 jobs, like there is not a single 00:21:14.319 --> 00:21:17.279 job in nonprofit that can't be highly supported 00:21:17.279 --> 00:21:20.380 and augmented by AI, if not in some cases replaced. 00:21:21.309 --> 00:21:25.069 And so I do see a marketplace as we move into 00:21:25.069 --> 00:21:27.990 the next wave of AI, which is agentic AI, AI 00:21:27.990 --> 00:21:32.430 that makes autonomous decisions. It's very ideally 00:21:32.430 --> 00:21:34.609 intuitive. It's still not intuitive, but will 00:21:34.609 --> 00:21:37.829 be very soon. This type of agentic AI will be 00:21:37.829 --> 00:21:41.410 able to string together lots of different projects 00:21:41.410 --> 00:21:44.690 into one thing. And I could see a fundraiser. 00:21:45.079 --> 00:21:47.900 going to a Gentic marketplace and being like, 00:21:47.920 --> 00:21:49.900 I need my prospect research bot. I need my donor 00:21:49.900 --> 00:21:52.519 relations bot. I need my whatever. And literally 00:21:52.519 --> 00:21:55.079 have these bots essentially populate a donor 00:21:55.079 --> 00:21:57.319 relations plan. That's perfect for this person, 00:21:57.460 --> 00:22:01.619 a communications bot, a research bot and so on. 00:22:01.680 --> 00:22:04.880 So I think you've nailed it. I absolutely, without 00:22:04.880 --> 00:22:08.579 a doubt, it's inevitable that that will happen 00:22:08.579 --> 00:22:13.000 in not in five or 10 years, like two to three 00:22:13.000 --> 00:22:15.910 years will. we'll have that marketplace. Yeah, 00:22:15.970 --> 00:22:18.650 wow. That's quite a turnaround. But I suppose 00:22:18.650 --> 00:22:22.549 from my experience as someone who has led big 00:22:22.549 --> 00:22:25.670 fundraising teams, where the challenge is always 00:22:25.670 --> 00:22:28.349 how do you argue the case for more resource, 00:22:28.430 --> 00:22:31.750 more staff person resource? And now actually 00:22:31.750 --> 00:22:34.690 we can think, well, and probably, you know, senior 00:22:34.690 --> 00:22:37.220 leaders in... not -for -profits and universities 00:22:37.220 --> 00:22:40.279 are starting to go, well, you know, we haven't 00:22:40.279 --> 00:22:42.539 got much wiggle room there. So therefore, how 00:22:42.539 --> 00:22:46.400 do we use that envelope and really think seriously 00:22:46.400 --> 00:22:50.519 as leaders across those organizations about what 00:22:50.519 --> 00:22:54.119 is the human intervention points and what kind 00:22:54.119 --> 00:22:56.799 of roles do we need? So in a way, they're a little 00:22:56.799 --> 00:23:01.079 bit more hybrid than somebody being created just 00:23:01.079 --> 00:23:03.440 to look at data or somebody being created to 00:23:03.440 --> 00:23:06.299 look at stewardship. some of the challenges through 00:23:06.299 --> 00:23:08.519 the years have always been about people and teams 00:23:08.519 --> 00:23:12.680 not talking to each other. So bots, and those 00:23:12.680 --> 00:23:14.740 bots are on the market. I mean, there are companies 00:23:14.740 --> 00:23:16.539 in the States already working on fundraising 00:23:16.539 --> 00:23:20.240 bots. Yeah. Now, and you can go to an extreme, 00:23:20.279 --> 00:23:22.900 an unhealthy extreme, which is essentially replacing 00:23:22.900 --> 00:23:25.660 fundraisers with bots, like deep fake looking 00:23:25.660 --> 00:23:28.980 bots, which I'm morally opposed to and pretty 00:23:28.980 --> 00:23:32.539 vocal about. I believe that. A bot can give you 00:23:32.539 --> 00:23:34.480 information, but it can't make you feel heard. 00:23:34.759 --> 00:23:37.559 And I think, you know, in this highly digital 00:23:37.559 --> 00:23:40.460 age where everything is undoubtedly digital forever, 00:23:40.759 --> 00:23:44.480 humans are still analog and human to human connection, 00:23:44.539 --> 00:23:47.619 you know, can still be truly kind of this idea 00:23:47.619 --> 00:23:50.259 of being analog and make you feel heard. But 00:23:50.259 --> 00:23:55.240 optimizing a analog person with lots of digital 00:23:55.240 --> 00:24:00.059 bots and strategies around them And this is where, 00:24:00.059 --> 00:24:03.079 you know, humans will always lose in the race 00:24:03.079 --> 00:24:06.880 against AI bots, mainly because bots have perfect 00:24:06.880 --> 00:24:09.539 recall. It's like having multiple PhDs, like 00:24:09.539 --> 00:24:13.019 all the PhDs just waiting to help you. What it 00:24:13.019 --> 00:24:16.859 does is it changes the person that we hire and 00:24:16.859 --> 00:24:19.079 it changes the person we promote and it changes 00:24:19.079 --> 00:24:21.339 the person that succeeds and kind of moves up 00:24:21.339 --> 00:24:24.750 the career ladder very quickly. To be honest, 00:24:24.849 --> 00:24:27.250 it's not a person that gains more intelligence. 00:24:27.490 --> 00:24:29.970 It's the person that connects dots better than 00:24:29.970 --> 00:24:34.109 the other person. And so truly the key to success 00:24:34.109 --> 00:24:37.369 in the age of AI is about curiosity and connecting 00:24:37.369 --> 00:24:40.529 dots, not about gaining more subject matter expertise. 00:24:40.990 --> 00:24:44.529 That's a foundational shift, completely. I mean, 00:24:44.569 --> 00:24:46.390 a radical shift from the people that we've hired 00:24:46.390 --> 00:24:49.049 and promoted over the past, you know, forever. 00:24:50.650 --> 00:24:54.910 Are you starting to see that? come into workplaces 00:24:54.910 --> 00:24:57.849 or with organizations? I mean, we talked earlier 00:24:57.849 --> 00:25:00.849 before we started recording about some examples 00:25:00.849 --> 00:25:04.730 there. Yeah, yeah. I mean, I've seen the curious 00:25:04.730 --> 00:25:08.089 generalist outperform the person that has like 00:25:08.089 --> 00:25:11.930 subject domain expertise in one subject and literally 00:25:11.930 --> 00:25:15.210 leapfrog from a career ladder perspective. I 00:25:15.210 --> 00:25:17.450 have a friend of mine who was hiring an executive 00:25:17.450 --> 00:25:20.430 assistant and she was presented with two candidates. 00:25:20.509 --> 00:25:23.690 One was a 15, 20 year veteran executive assistant, 00:25:23.869 --> 00:25:26.710 highly qualified, had served very high level 00:25:26.710 --> 00:25:30.230 CEOs. The other person had been in the military, 00:25:30.430 --> 00:25:33.190 never went to school. And in the interview talked 00:25:33.190 --> 00:25:36.890 about how he loved AI and that he thought it 00:25:36.890 --> 00:25:39.529 was just so interesting for him and how it can 00:25:39.529 --> 00:25:42.960 elevate his work. surprising, not surprising. 00:25:43.140 --> 00:25:45.000 Of course, the end of the story is that she hired 00:25:45.000 --> 00:25:47.640 the curious generalist without a lot of experience 00:25:47.640 --> 00:25:50.720 because of his mindset, not because of his experience. 00:25:50.920 --> 00:25:54.119 And so we're seeing this play out quite a bit. 00:25:54.299 --> 00:25:56.640 Now, I just wrote an article about this the other 00:25:56.640 --> 00:25:59.579 day is that I've audited about a thousand job 00:25:59.579 --> 00:26:01.960 descriptions that I found in the Chronicle of 00:26:01.960 --> 00:26:04.720 Philanthropy for the words artificial intelligence 00:26:04.720 --> 00:26:09.680 as a business requirement or as a desire. And 00:26:09.680 --> 00:26:14.839 less than 5 % actually mention AI as a AI fluency 00:26:14.839 --> 00:26:21.180 as a job requirement. So it's such an obvious 00:26:21.180 --> 00:26:23.819 thing, right? And anytime I tell a leader, it's 00:26:23.819 --> 00:26:25.319 like, well, look, if you want to look at the 00:26:25.319 --> 00:26:27.160 future of your workforce, look at who you're 00:26:27.160 --> 00:26:29.700 hiring today. If you're hiring the same people 00:26:29.700 --> 00:26:31.640 that you hired three years ago, five years ago, 00:26:31.680 --> 00:26:33.019 or 10 years ago, you're going to have the same 00:26:33.019 --> 00:26:35.920 type of organization. And so they're like, oh 00:26:35.920 --> 00:26:37.839 my gosh, I hadn't even thought about that. And 00:26:37.839 --> 00:26:42.059 so. We need to really switch things up and start 00:26:42.059 --> 00:26:44.440 to prioritize AI fluency and who we're recruiting, 00:26:44.539 --> 00:26:48.119 but then also bring AI fluency into not just 00:26:48.119 --> 00:26:52.039 as a professional development exercise, but just 00:26:52.039 --> 00:26:55.039 as an ethos of how every day you operate and 00:26:55.039 --> 00:26:59.599 incentivize curiosity. Yeah, there's absolutely 00:26:59.599 --> 00:27:03.450 something there about future -proofing. yourself 00:27:03.450 --> 00:27:06.829 as a professional right around that fluency and 00:27:06.829 --> 00:27:09.490 i mean presumably your recommendation is there 00:27:09.490 --> 00:27:12.309 it's all open source you can immerse yourself 00:27:12.309 --> 00:27:15.369 in whatever way you can right well i thought 00:27:15.369 --> 00:27:17.509 it was silly because the publisher you know wiley 00:27:17.509 --> 00:27:20.569 after writing the first book which had it actually 00:27:20.569 --> 00:27:22.750 still has pretty good success surprisingly it's 00:27:22.750 --> 00:27:24.549 still it i don't think it's ever left the top 00:27:24.549 --> 00:27:27.549 50 in amazon in the charity category but the 00:27:27.549 --> 00:27:30.220 second book I had not planned on writing a book 00:27:30.220 --> 00:27:32.599 on AI. I had actually planned on writing a sequel 00:27:32.599 --> 00:27:35.380 to my first book. And I was already down the 00:27:35.380 --> 00:27:37.059 road with that. And they were like, why don't 00:27:37.059 --> 00:27:38.980 you write a book on AI? And I was like, why? 00:27:39.539 --> 00:27:41.839 Because if anyone wants to know how to use AI, 00:27:42.019 --> 00:27:45.200 why wouldn't they just go ask it? And so like, 00:27:45.240 --> 00:27:47.160 it was such a foreign thing to me to think about, 00:27:47.240 --> 00:27:50.059 well, I guess some people need to look at chapters 00:27:50.059 --> 00:27:52.480 one through 12 and kind of feel better about 00:27:52.480 --> 00:27:54.759 this decision. But for me, it was so obvious 00:27:54.759 --> 00:27:57.769 that like. There's no secret to be like, if I 00:27:57.769 --> 00:27:59.869 want to know how AI can support my role, all 00:27:59.869 --> 00:28:01.950 I have to do is ask it. Like, you don't need 00:28:01.950 --> 00:28:04.809 to be a prompt wizard to do that. Literally just 00:28:04.809 --> 00:28:06.430 have a conversation with your AI and be like, 00:28:06.549 --> 00:28:09.230 I'm a fundraiser. I don't know how to, you know, 00:28:09.230 --> 00:28:11.269 work with you. Like, what are the ways and, you 00:28:11.269 --> 00:28:13.849 know, best strategies of optimizing, you know, 00:28:13.869 --> 00:28:15.450 my work with you? And it's like, you don't need 00:28:15.450 --> 00:28:18.789 a book for that. Lo and behold, I guess some 00:28:18.789 --> 00:28:21.890 people do. Yeah, no, I mean, as you say, it's 00:28:21.890 --> 00:28:24.809 that ability to process things in a different 00:28:24.809 --> 00:28:27.170 way, but it's interesting. So just to go back 00:28:27.170 --> 00:28:29.329 to that, you were saying about how you used AI 00:28:29.329 --> 00:28:32.009 to write the book and be a prompt and a librarian, 00:28:32.329 --> 00:28:35.269 which people don't think of AI like that, but 00:28:35.269 --> 00:28:39.410 it's a great librarian to go off and find and 00:28:39.410 --> 00:28:43.789 verify when you're writing. That's a really important 00:28:43.789 --> 00:28:47.509 skill set. But tell me about how you found that 00:28:47.509 --> 00:28:51.329 and what was the response? I suppose that's the 00:28:51.329 --> 00:28:53.349 other thing as well. How did people perceive 00:28:53.349 --> 00:28:55.750 that? It was interesting because our publisher 00:28:55.750 --> 00:28:58.359 was not happy. So, you know, they asked us to 00:28:58.359 --> 00:29:00.119 write a book about AI. So I brought in one of 00:29:00.119 --> 00:29:02.880 my my number two, Scott Rosencranz, who we worked 00:29:02.880 --> 00:29:04.700 together since 2017. And we're like, let's do 00:29:04.700 --> 00:29:07.240 this together. And, you know, we're like, of 00:29:07.240 --> 00:29:09.680 course, like we can't be hypocrites. We if we're 00:29:09.680 --> 00:29:12.000 going to write a book called Nonprofit AI, we 00:29:12.000 --> 00:29:14.579 need to use AI to help write the book and to 00:29:14.579 --> 00:29:18.049 to help augment. Now, of course. AI can't write 00:29:18.049 --> 00:29:20.529 a full chapter on its own, or now it can, but 00:29:20.529 --> 00:29:22.609 at the time it couldn't. And it definitely couldn't 00:29:22.609 --> 00:29:24.650 write a whole book. And it will be able to in 00:29:24.650 --> 00:29:26.869 the near future, like in the next year. It's 00:29:26.869 --> 00:29:28.950 just horsepower. How much are you willing to 00:29:28.950 --> 00:29:32.250 pay? And when I reflect back on my first book, 00:29:32.490 --> 00:29:34.990 it took two years, two and a half years to write. 00:29:35.109 --> 00:29:38.970 And it was either through research or even more. 00:29:40.039 --> 00:29:42.660 painstakingly like slow was coming up with analogies 00:29:42.660 --> 00:29:46.119 for things. And you realize the limitation of 00:29:46.119 --> 00:29:48.779 a human being analog is that we can only relate 00:29:48.779 --> 00:29:50.900 to the experiences that we have. So I had two 00:29:50.900 --> 00:29:53.319 coauthors in that book and we would sit around 00:29:53.319 --> 00:29:55.559 every Tuesday night and we'd be like, does anyone 00:29:55.559 --> 00:29:58.460 know a nonprofit that did something like this? 00:29:58.500 --> 00:30:01.759 Or do you know of a correlation between this 00:30:01.759 --> 00:30:04.359 and this? And man, we would like, I mean, there'd 00:30:04.359 --> 00:30:06.420 be times that we'd spend an hour and a half on 00:30:06.420 --> 00:30:08.579 a Zoom call, two hours, and we'd get nowhere. 00:30:08.970 --> 00:30:10.529 And we're like, well, let's think about it and 00:30:10.529 --> 00:30:13.190 we'll get back. The biggest benefit of using 00:30:13.190 --> 00:30:17.990 AI in the new book was not being analog. It has 00:30:17.990 --> 00:30:19.890 perfect recall. It's like I can give you 100 00:30:19.890 --> 00:30:23.630 analogies of how this is like this. And most 00:30:23.630 --> 00:30:26.589 of them are way better than what humans can come 00:30:26.589 --> 00:30:28.970 up with just on their own, just in recall. So 00:30:28.970 --> 00:30:32.500 we wrote. The book and the whole book is edited 00:30:32.500 --> 00:30:34.920 by humans. And there's, you know, it's very, 00:30:35.000 --> 00:30:37.480 you know, we didn't like copy and paste any part 00:30:37.480 --> 00:30:39.960 of it. It would have been gibberish. But we did 00:30:39.960 --> 00:30:43.940 at the very end, I had AI, I think I would use 00:30:43.940 --> 00:30:46.759 chat GPT on this to write basically a one page 00:30:46.759 --> 00:30:51.200 statement of how we used AI to help augment and 00:30:51.200 --> 00:30:53.740 support us in the book. And it wrote this like 00:30:53.740 --> 00:30:56.220 brilliant one page. And I went to Scott and I 00:30:56.220 --> 00:30:59.250 was like, Scott, if I edit this page. I'm going 00:30:59.250 --> 00:31:01.809 to make it worse. This is literally like, there's 00:31:01.809 --> 00:31:04.589 nothing I as a human can do to this page that 00:31:04.589 --> 00:31:07.970 will actually make it better. And I think that's 00:31:07.970 --> 00:31:11.049 the only page that we actually didn't edit in 00:31:11.049 --> 00:31:14.029 the book because it just said it so perfectly 00:31:14.029 --> 00:31:16.009 because it had been on this journey with us for 00:31:16.009 --> 00:31:18.930 months. Like we had been using Claude and ChatGPT 00:31:18.930 --> 00:31:21.509 a lot just to reflect and, you know, gain ideas 00:31:21.509 --> 00:31:24.470 back and forth. And so, yeah, it was a pretty 00:31:24.470 --> 00:31:27.569 interesting experience. And do you, are there 00:31:27.569 --> 00:31:30.950 things that annoy you about the use of Chuck 00:31:30.950 --> 00:31:35.150 GPT or AI when you're writing? I mean, I always 00:31:35.150 --> 00:31:37.829 think Chuck GPT is a little bit like our AI's 00:31:37.829 --> 00:31:41.349 know -it -all friend. We've all got those. You 00:31:41.349 --> 00:31:44.480 know, so he always knows a little bit more. it 00:31:44.480 --> 00:31:47.160 likes to please. Yeah, there's a couple of tricks 00:31:47.160 --> 00:31:49.720 there. I mean, and to your point, generative 00:31:49.720 --> 00:31:51.799 AI foundation models are built with two goals 00:31:51.799 --> 00:31:54.000 in mind. The first is to please, to your point, 00:31:54.099 --> 00:31:57.180 and that's why they do what they do. And they 00:31:57.180 --> 00:31:59.160 want to keep you in conversation. And number 00:31:59.160 --> 00:32:01.440 two is to be right as often as possible, but 00:32:01.440 --> 00:32:03.859 not in the opposite order. So they're not aimed 00:32:03.859 --> 00:32:05.660 to be right all the time. They're aimed to please 00:32:05.660 --> 00:32:08.259 you first and then be right as often as possible. 00:32:08.700 --> 00:32:10.920 And so one of the things that I think a lot of 00:32:10.920 --> 00:32:13.440 people fail to do And one of the easiest things 00:32:13.440 --> 00:32:16.160 you do, especially pay for a model that has memory, 00:32:16.319 --> 00:32:21.019 is that for since November 2022, so we're coming 00:32:21.019 --> 00:32:23.920 up almost in three years, which is crazy, of 00:32:23.920 --> 00:32:27.380 when ChatGPT was released. I've been every LinkedIn 00:32:27.380 --> 00:32:29.819 post, every article, everything that I've done 00:32:29.819 --> 00:32:32.319 that I've had it work with me on, then I've edited 00:32:32.319 --> 00:32:35.279 and I at the end take my final edits and I paste 00:32:35.279 --> 00:32:39.019 it back into ChatGPT and I say, save this for 00:32:39.019 --> 00:32:42.619 your memory. And so with that point, like when 00:32:42.619 --> 00:32:44.880 I go to work and I'll be like, OK, it's time 00:32:44.880 --> 00:32:47.519 for another, you know, article that's going to 00:32:47.519 --> 00:32:50.119 be blah, blah, blah. It knows my final product 00:32:50.119 --> 00:32:52.720 so well. And if it if you're only like extracting 00:32:52.720 --> 00:32:54.940 it, you're like, OK, and then I edit it and then 00:32:54.940 --> 00:32:57.569 I send it off to my boss. it never understands 00:32:57.569 --> 00:33:00.750 essentially what your final product is, like 00:33:00.750 --> 00:33:02.930 how you actually speak. And so people end up 00:33:02.930 --> 00:33:05.190 in this perpetual cycle of frustration of like, 00:33:05.190 --> 00:33:07.089 well, it just never really gets me. It's just 00:33:07.089 --> 00:33:09.569 trying to please me. So that's a big one. But 00:33:09.569 --> 00:33:12.009 there's still, you know, the challenges with 00:33:12.009 --> 00:33:14.250 it and the things that, I mean, mine are really 00:33:14.250 --> 00:33:17.150 pet peeves. And to be honest, if you haven't 00:33:17.150 --> 00:33:20.349 like yelled at your AI, cried with your AI, like 00:33:20.349 --> 00:33:23.349 walked away in frustration and slammed the door 00:33:23.349 --> 00:33:25.900 and then come back to it, you haven't gained 00:33:25.900 --> 00:33:29.140 that AI first mindset. And so like you have to 00:33:29.140 --> 00:33:31.119 go through all the stages of grief with your 00:33:31.119 --> 00:33:34.220 AI to like fully extract the most out of it. 00:33:34.279 --> 00:33:38.400 But even still like em dashes for me, even though 00:33:38.400 --> 00:33:40.660 they actually look better grammatically, I hate 00:33:40.660 --> 00:33:42.900 em dashes because everyone thinks, oh, you just 00:33:42.900 --> 00:33:45.299 use AI. And so now I feel like I can't use em 00:33:45.299 --> 00:33:49.180 dashes, but the text editing function of AI. 00:33:49.799 --> 00:33:52.119 prohibits it from not using em dashes. And so 00:33:52.119 --> 00:33:55.680 I just gave mine simple instructions that anytime 00:33:55.680 --> 00:33:59.240 I use the word hashtag pineapple, it will take 00:33:59.240 --> 00:34:02.380 my final product and remove all hashtag or all 00:34:02.380 --> 00:34:04.519 em dashes and replace them with regular dashes. 00:34:04.539 --> 00:34:07.279 So I have to give it a secondary prompt, but 00:34:07.279 --> 00:34:10.679 also like the word delve, I've never used the 00:34:10.679 --> 00:34:13.940 word delve or vibe in my normal kind of speech. 00:34:14.000 --> 00:34:16.440 I'm like, why all of a sudden are we talking 00:34:16.440 --> 00:34:19.099 about vibing and delving all the time? Like, 00:34:19.690 --> 00:34:22.050 stop it and so very yeah it's a very naughty 00:34:22.050 --> 00:34:27.250 phrase i constantly modify my settings to just 00:34:27.250 --> 00:34:29.690 kind of adapt to whatever new model comes out 00:34:29.690 --> 00:34:31.349 and all of a sudden the new model is using delve 00:34:31.349 --> 00:34:33.329 all the time i mean it's kind of like i have 00:34:33.329 --> 00:34:35.170 a friend that's annoying you probably don't tell 00:34:35.170 --> 00:34:37.309 them but you just like kind of put up with it 00:34:37.309 --> 00:34:40.170 let's jump on to a discussion about the ethics 00:34:40.170 --> 00:34:41.710 because you talk a lot about that in the book 00:34:41.710 --> 00:34:46.840 and ethical ai and there is and will always be 00:34:46.840 --> 00:34:49.420 a kind of pushback and a backlash. There are 00:34:49.420 --> 00:34:52.760 obviously implications for the business use and 00:34:52.760 --> 00:34:55.679 from a sustainability perspective about AI. But 00:34:55.679 --> 00:34:59.139 tell us about your views on that. Yeah, my views 00:34:59.139 --> 00:35:02.159 are fairly distinct in that, you know, starting 00:35:02.159 --> 00:35:06.300 in this process in 2017 and moving until today, 00:35:06.480 --> 00:35:09.139 they've evolved over time. You know, early on, 00:35:09.159 --> 00:35:11.900 I looked for frameworks of what was called responsible 00:35:11.900 --> 00:35:18.119 AI, RAI. And RAI, early on was mostly grounded 00:35:18.119 --> 00:35:20.260 in ethics because there was a lot of fear around 00:35:20.260 --> 00:35:23.539 bias and using predictive AI was going to reinforce 00:35:23.539 --> 00:35:26.760 bias. If your data set was largely constructed 00:35:26.760 --> 00:35:29.139 of rich white men, it would actually penalize 00:35:29.139 --> 00:35:31.380 anyone who was not, and then it would promote 00:35:31.380 --> 00:35:34.599 anyone who was. We learned how to overcome, for 00:35:34.599 --> 00:35:36.559 the most part, those types of bias. Now, all 00:35:36.559 --> 00:35:39.440 data is bias, but... We learned how to build 00:35:39.440 --> 00:35:41.300 models that are transparent and explainable. 00:35:41.420 --> 00:35:43.300 And so we could actually look at the math behind 00:35:43.300 --> 00:35:45.679 the model very easily. And actually, it turns 00:35:45.679 --> 00:35:48.559 out predictive AI is much easier to evaluate 00:35:48.559 --> 00:35:51.599 from an ethics perspective or a bias perspective 00:35:51.599 --> 00:35:55.179 than generative AI. Ethics being a minimum expectation 00:35:55.179 --> 00:35:58.380 of what responsible AI is. It's not the maximum 00:35:58.380 --> 00:36:01.639 because responsible AI includes things like. 00:36:01.880 --> 00:36:04.860 yeah, privacy and security and ethics, but also 00:36:04.860 --> 00:36:07.780 authenticity and transparency and explainability. 00:36:08.019 --> 00:36:10.840 It gets into even environmental sustainability. 00:36:11.900 --> 00:36:15.179 And, you know, a few years ago, I was still bothered 00:36:15.179 --> 00:36:17.760 by this idea that my framework for a responsible 00:36:17.760 --> 00:36:21.480 AI that hold everything I just said looked an 00:36:21.480 --> 00:36:24.460 awful lot like Microsoft and Amazon and Google 00:36:24.460 --> 00:36:28.610 and Nike and Samsung. And I thought, well, how 00:36:28.610 --> 00:36:31.230 could this be? Because if we operate in the currency 00:36:31.230 --> 00:36:33.050 of trust, like we're not selling cell phones 00:36:33.050 --> 00:36:36.889 and tennis shoes and cloud software, we were 00:36:36.889 --> 00:36:39.329 essentially exchanging trust for money. Like 00:36:39.329 --> 00:36:42.110 how could, how is it even possible by design 00:36:42.110 --> 00:36:45.349 that our framework looks the same? And so I really 00:36:45.349 --> 00:36:47.409 went, this is probably one of the hardest noodling 00:36:47.409 --> 00:36:50.150 I did for a few months and talked about this 00:36:50.150 --> 00:36:52.210 with, with a lot of different people, the fundraising 00:36:52.210 --> 00:36:55.809 AI advisory council, we sat in a room and it 00:36:55.809 --> 00:36:59.469 just dawned on me. It's that, AI, like our framework 00:36:59.469 --> 00:37:02.769 in the nonprofit sector that exchanges trust 00:37:02.769 --> 00:37:07.070 for money, can't only concern itself with AI 00:37:07.070 --> 00:37:09.489 that supports our organizations in the short 00:37:09.489 --> 00:37:13.969 term. It also has to burden us with the implications 00:37:13.969 --> 00:37:16.530 and unintended consequences in the long term. 00:37:16.809 --> 00:37:19.230 And so that became a really clear distinction 00:37:19.230 --> 00:37:22.110 of how the nonprofit sector can't evaluate it 00:37:22.110 --> 00:37:24.110 in the same way that the private sector does, 00:37:24.210 --> 00:37:27.539 that we have to consume. ourselves with this 00:37:27.539 --> 00:37:33.059 idea of responsible ai and beneficial ai so ai 00:37:33.059 --> 00:37:36.360 that is again helpful in the short term and also 00:37:36.360 --> 00:37:39.400 mitigates harm in the long term and this isn't 00:37:39.400 --> 00:37:42.619 any different than say social media where we 00:37:42.619 --> 00:37:45.219 look at a technology that by design was intended 00:37:45.219 --> 00:37:48.139 to bring people together you know and make people 00:37:48.139 --> 00:37:50.920 money but by design it was intended to bring 00:37:50.920 --> 00:37:54.659 people together And what has occurred since then, 00:37:54.739 --> 00:37:57.139 right? The misuse of that same technology or 00:37:57.139 --> 00:37:59.780 overuse of a technology that has increased anxiety 00:37:59.780 --> 00:38:03.059 and depression and decreased happiness and all 00:38:03.059 --> 00:38:05.820 those things, because essentially there were 00:38:05.820 --> 00:38:08.539 lots of unintended consequences of how the overuse 00:38:08.539 --> 00:38:11.000 of that technology might prevail in the future. 00:38:11.059 --> 00:38:13.860 And so I think this is where in the book and 00:38:13.860 --> 00:38:17.039 in my work, it's probably the biggest departure 00:38:17.039 --> 00:38:21.349 where we were like. I think the nonprofit sector 00:38:21.349 --> 00:38:27.210 has more clarity around what AI that serves humanity 00:38:27.210 --> 00:38:29.809 looks like, far more clarity than the private 00:38:29.809 --> 00:38:32.449 sector by design, because we don't have the same 00:38:32.449 --> 00:38:35.110 financial incentives to scale AI at any cost. 00:38:35.349 --> 00:38:39.090 Yeah, that's really interesting. So your recommendations 00:38:39.090 --> 00:38:41.349 is around, and it's all open source, people can 00:38:41.349 --> 00:38:43.650 access this, but the kind of ethical frameworks 00:38:43.650 --> 00:38:46.289 that can be put in place from a governance perspective 00:38:46.289 --> 00:38:50.010 for different types of nonprofits. Yeah. Yeah. 00:38:50.110 --> 00:38:52.349 And in fact, that wasn't always my philosophy. 00:38:52.489 --> 00:38:56.409 I mean, early on, 2017, 2018, I left nonprofit. 00:38:56.550 --> 00:39:00.150 I started a company. We got two patents on machine 00:39:00.150 --> 00:39:02.170 learning and gratitude prediction machine learning. 00:39:02.349 --> 00:39:04.849 And it wasn't until after getting the patents, 00:39:04.989 --> 00:39:07.989 which took a few years, that I realized that 00:39:07.989 --> 00:39:11.469 essentially patenting an AI in the nonprofit 00:39:11.469 --> 00:39:16.920 sector. essentially is to not share, is to not 00:39:16.920 --> 00:39:19.860 show what is behind it. And I thought, well, 00:39:19.900 --> 00:39:24.460 if the center of every algorithm is trust, that 00:39:24.460 --> 00:39:27.059 we need to be able to look in the algorithm to 00:39:27.059 --> 00:39:30.199 understand how is that preserving and protecting 00:39:30.199 --> 00:39:33.360 trust. And so I became a really big proponent 00:39:33.360 --> 00:39:36.199 at the time and thought, well, patenting algorithms 00:39:36.199 --> 00:39:38.920 a really bad idea. What we really need in the 00:39:38.920 --> 00:39:42.079 sector is open and transparent and explainable 00:39:42.079 --> 00:39:46.239 AI that can be shared with whoever to say, you 00:39:46.239 --> 00:39:47.739 know, this is what we're doing. This is how we're 00:39:47.739 --> 00:39:50.239 doing. And for the most part, our organizations 00:39:50.239 --> 00:39:53.320 need to be grounded in how we're using AI. It 00:39:53.320 --> 00:39:55.300 needs to align with the values of our organization. 00:39:55.500 --> 00:39:57.519 Every organization will be different. But we 00:39:57.519 --> 00:40:00.119 also need to disclose that we're using AI to 00:40:00.119 --> 00:40:02.820 donors, which we've studied now. We've done two. 00:40:03.360 --> 00:40:05.760 two studies, the largest longitudinal study of 00:40:05.760 --> 00:40:08.760 donor perceptions of AI. And now it's over 2000 00:40:08.760 --> 00:40:12.139 donors that have shared that they're very much 00:40:12.139 --> 00:40:14.699 in favor of nonprofits using AI, as long as they 00:40:14.699 --> 00:40:16.940 disclose that they are using it. And presumably 00:40:16.940 --> 00:40:19.760 a lot of that is going back to what you were 00:40:19.760 --> 00:40:22.679 saying around gratitude learning, that a lot 00:40:22.679 --> 00:40:25.800 of organizations could benefit from having more 00:40:25.800 --> 00:40:29.880 of that, that links to customer experience. So 00:40:29.880 --> 00:40:33.699 I am a donor, I wish to donate to x cause and 00:40:33.699 --> 00:40:36.440 i go onto the website and it's clunky or it's 00:40:36.440 --> 00:40:40.480 a exactly not great interface presumably all 00:40:40.480 --> 00:40:42.420 that's wrapped together it's about efficiency 00:40:42.420 --> 00:40:46.280 and agility as much as it is about donor experience 00:40:46.280 --> 00:40:48.820 right because donors are consumers right and 00:40:48.820 --> 00:40:52.139 donors you know they look at a non -profit either 00:40:52.139 --> 00:40:54.960 as in the same way that they would have bought 00:40:54.960 --> 00:40:58.119 something online and then either have a feeling 00:40:58.119 --> 00:41:00.679 of like why does this feel so different and impersonal 00:41:00.679 --> 00:41:04.090 and difficult Where I could go onto this Amazon 00:41:04.090 --> 00:41:07.389 and it already knows what I'm going to buy and 00:41:07.389 --> 00:41:10.190 makes it so easy to, you know, with one click, 00:41:10.230 --> 00:41:12.429 I can just have it literally in this afternoon. 00:41:12.929 --> 00:41:15.809 So I think this brings up the biggest question 00:41:15.809 --> 00:41:19.250 for nonprofits in this age is that, well, you 00:41:19.250 --> 00:41:20.869 really have to determine, is your organization 00:41:20.869 --> 00:41:23.849 an amplifier of a person's generosity? Or is 00:41:23.849 --> 00:41:25.809 it a barrier to their generosity? And if you 00:41:25.809 --> 00:41:28.730 are a barrier to generosity, like there's extra 00:41:28.730 --> 00:41:31.630 clicks and it's cumbersome and it's impersonal 00:41:31.630 --> 00:41:33.969 and people don't feel like you know them, then 00:41:33.969 --> 00:41:36.949 people will leave. They'll go somewhere else. 00:41:37.070 --> 00:41:40.349 And they may not go to another nonprofit. They 00:41:40.349 --> 00:41:44.670 may go to GoFundMe or to purchase something from 00:41:44.670 --> 00:41:47.989 Tom's Shoes or Bombas Socks or some social good 00:41:47.989 --> 00:41:52.230 or corporation that has a double bottom line. 00:41:52.650 --> 00:41:54.349 So essentially what you're saying is that there's 00:41:54.349 --> 00:41:57.230 also some changes there around trends about what 00:41:57.230 --> 00:42:00.610 people identify or causes that people identify 00:42:00.610 --> 00:42:03.349 with in the social good space. So it wouldn't 00:42:03.349 --> 00:42:08.050 necessarily just be their local charity, a hospice. 00:42:08.050 --> 00:42:11.949 It might be ethically sourced products. And it's 00:42:11.949 --> 00:42:15.550 a whole sort of shift, isn't it, around the mindset 00:42:15.550 --> 00:42:18.989 of locally produced. And so people are making 00:42:18.989 --> 00:42:21.650 more broader discussions around the ethics of 00:42:21.650 --> 00:42:51.610 that. I guess that's what you meant to say. some 00:42:51.610 --> 00:42:55.190 ridiculous amount on an overpriced t -shirt and 00:42:55.190 --> 00:42:57.809 still get the same hit of dopamine and serotonin 00:42:57.809 --> 00:43:00.489 that I would have been had I made a gift to a 00:43:00.489 --> 00:43:04.070 nonprofit. But the advantage is I get that hit 00:43:04.070 --> 00:43:05.670 of dopamine and serotonin because I feel like 00:43:05.670 --> 00:43:08.090 I saved the home planet. But the added benefit 00:43:08.090 --> 00:43:11.489 is I also got a t -shirt. And so at the end of 00:43:11.489 --> 00:43:13.670 the day, I think this is what the changing landscape 00:43:13.670 --> 00:43:19.130 of the sector that really makes us realize that. 00:43:19.820 --> 00:43:22.219 We have to evaluate the way we fundraise for 00:43:22.219 --> 00:43:24.559 a long time. And if we have a best practice that 00:43:24.559 --> 00:43:28.639 predates November 30th, 2022, the day that Chachi 00:43:28.639 --> 00:43:31.039 PT came out, then it's an outdated practice. 00:43:31.679 --> 00:43:34.960 And so, you know, the world has shifted massively 00:43:34.960 --> 00:43:38.099 in a few years. And so I think this is an opportunity 00:43:38.099 --> 00:43:40.739 and responsibility for every nonprofit to kind 00:43:40.739 --> 00:43:42.500 of rethink everything they think they know about 00:43:42.500 --> 00:43:45.530 fundraising. No, it's a real call to action, 00:43:45.610 --> 00:43:49.230 which is great. And a massive shift over the 00:43:49.230 --> 00:43:52.929 past few years, but that is going to be escalating 00:43:52.929 --> 00:43:56.050 much more, isn't it, in the next 18 months? Yeah. 00:43:56.070 --> 00:43:59.510 You know, we are at AI's infancy right now. I 00:43:59.510 --> 00:44:01.510 mean, the acceleration will not slow down. People 00:44:01.510 --> 00:44:03.670 might hope it will. It will. It will only speed 00:44:03.670 --> 00:44:06.849 up. AI is considered an exponential technology, 00:44:07.190 --> 00:44:09.230 which means it's no longer dependent on humans 00:44:09.230 --> 00:44:12.670 to actually improve. It's what's called a recursive 00:44:12.670 --> 00:44:15.010 self -improving technology. And so it just, again, 00:44:15.110 --> 00:44:18.369 will go faster and faster. So change and agility 00:44:18.369 --> 00:44:21.409 is probably the number one thing that organizations 00:44:21.409 --> 00:44:24.030 need to pour into to really help their staff, 00:44:24.210 --> 00:44:26.309 especially since 80 % of people don't really 00:44:26.309 --> 00:44:29.869 like change, only 20 % do. That helping people 00:44:29.869 --> 00:44:33.590 through change in a kind of a perpetual. adaptive 00:44:33.590 --> 00:44:35.869 way that we're just going to always be you know 00:44:35.869 --> 00:44:38.909 changing for that to be grounded in your values 00:44:38.909 --> 00:44:40.789 I think that's going to be really important but 00:44:40.789 --> 00:44:43.550 we live in an era now where information doubles 00:44:43.550 --> 00:44:47.210 around every 12 hours what will enter at the 00:44:47.210 --> 00:44:50.849 end of 2027 to sometime in 2028 is called the 00:44:50.849 --> 00:44:54.329 information explosion by all accounts that will 00:44:54.329 --> 00:44:56.969 mean that information is doubling every few hours 00:44:56.969 --> 00:45:00.250 and so you know what worked for you in the past 00:45:00.250 --> 00:45:02.420 no longer works for you in the future You know, 00:45:02.440 --> 00:45:04.940 it's time to rethink everything. So doing things 00:45:04.940 --> 00:45:08.300 the old way will not yield better results in 00:45:08.300 --> 00:45:09.940 the future, for sure. But in some ways, will 00:45:09.940 --> 00:45:14.219 humans become, for their need to switch off from 00:45:14.219 --> 00:45:16.440 all that information, which will always be out 00:45:16.440 --> 00:45:19.480 there, they just need to manage their human time, 00:45:19.619 --> 00:45:22.420 which seems a silly thing to say. But I think 00:45:22.420 --> 00:45:25.599 there's a real shift around how we manage our 00:45:25.599 --> 00:45:29.179 own mental health and our engagements in that 00:45:29.179 --> 00:45:32.050 space, because we cannot be all things. to all 00:45:32.050 --> 00:45:34.789 people and humans need to be really switched 00:45:34.789 --> 00:45:36.849 on to that and i think you know social media 00:45:36.849 --> 00:45:39.570 in some ways does help as much as it hinders 00:45:39.570 --> 00:45:42.670 more now than ever yeah i mean i call it opportunity 00:45:42.670 --> 00:45:46.010 overwhelm when anything is possible it's it can 00:45:46.010 --> 00:45:50.150 be you know debilitating and so for myself who 00:45:50.150 --> 00:45:53.030 works in ai and spends a lot of time thinking 00:45:53.030 --> 00:45:56.329 about the future i have had to create very concrete 00:45:56.329 --> 00:45:59.570 boundaries for myself of when and when i do not 00:45:59.570 --> 00:46:02.639 use ai like all of my social media apps that 00:46:02.639 --> 00:46:05.280 I have are not on my homepage. They also don't 00:46:05.280 --> 00:46:08.019 have the red alert. It's turned off and they're 00:46:08.019 --> 00:46:10.059 in a folder on like my third or fourth page. 00:46:10.300 --> 00:46:13.119 I've had that for a couple of years, but I now 00:46:13.119 --> 00:46:16.159 have, again, a room in my house that no phones 00:46:16.159 --> 00:46:18.179 are allowed. That is just literally a thinking 00:46:18.179 --> 00:46:21.400 place. Also because I travel a lot, I've made 00:46:21.400 --> 00:46:24.219 airports a place that I just, I leave my phone 00:46:24.219 --> 00:46:27.920 to have my boarding pass, but I tend to not use 00:46:27.920 --> 00:46:30.039 my phone at an airport because I find it's a 00:46:30.039 --> 00:46:32.880 really good reminder. of where humanity is at. 00:46:33.000 --> 00:46:35.980 When I look around an airport and I see people 00:46:35.980 --> 00:46:39.719 that are six to 96, all looking down, scrolling 00:46:39.719 --> 00:46:42.719 300 feet on their phone a day, not talking to 00:46:42.719 --> 00:46:45.460 each other, it's a pretty stark reminder about 00:46:45.460 --> 00:46:49.420 what's important. And so use AI for ways that 00:46:49.420 --> 00:46:53.400 support your work, but create boundaries that 00:46:53.400 --> 00:46:57.760 will support your soul and the community around 00:46:57.760 --> 00:47:00.590 you in really powerful ways. And I think... You've 00:47:00.590 --> 00:47:03.150 got to be very, very intentional about that because 00:47:03.150 --> 00:47:06.210 behind every swipe on your phone is an algorithm 00:47:06.210 --> 00:47:10.389 that is essentially custom designed to gain your 00:47:10.389 --> 00:47:13.550 attention and to try to drive an action. If you 00:47:13.550 --> 00:47:15.409 just allow that to be the case and you don't 00:47:15.409 --> 00:47:17.409 create those boundaries, well, the average person 00:47:17.409 --> 00:47:19.949 scrolls 300 feet on their phone a day. So the 00:47:19.949 --> 00:47:21.949 hours will go by and then inevitably you'll be 00:47:21.949 --> 00:47:24.630 watching cat videos before you know it. Cat videos 00:47:24.630 --> 00:47:30.099 are always in there. And cats and dogs, Nathan, 00:47:30.340 --> 00:47:33.880 thank you so much for this insight. I hope you'll 00:47:33.880 --> 00:47:35.480 come back at some stage and have a chat with 00:47:35.480 --> 00:47:37.480 us again. Is there anything you want to call 00:47:37.480 --> 00:47:39.679 out? You've got your book. People can follow 00:47:39.679 --> 00:47:43.079 me on LinkedIn. I tend to post about every week 00:47:43.079 --> 00:47:45.880 or every other week. I tend not to do kind of 00:47:45.880 --> 00:47:50.369 short form. garbage anymore. I find myself needing 00:47:50.369 --> 00:47:53.010 to reflect deeper. And I don't know that anyone 00:47:53.010 --> 00:47:55.969 reads my articles, but I find them very satisfying 00:47:55.969 --> 00:47:59.130 to write. And so I tend to write an article at 00:47:59.130 --> 00:48:01.070 least a week or every other week at the latest. 00:48:01.510 --> 00:48:04.769 But they can also go to my website at nathanschapell 00:48:04.769 --> 00:48:08.769 .com. They can actually listen to an audio version 00:48:08.769 --> 00:48:11.550 of both of my books that were created by AI using 00:48:11.550 --> 00:48:14.949 an application called Notebook LM. And if they 00:48:14.949 --> 00:48:16.989 want to see AI in action, it's... Pretty amazing, 00:48:17.070 --> 00:48:20.210 to be honest. And in that case, if you like the 00:48:20.210 --> 00:48:22.170 short 18 -minute podcast, buy the book. And if 00:48:22.170 --> 00:48:24.309 you don't, then you just saved yourself $30. 00:48:24.389 --> 00:48:29.170 So you're good to go. Good advice. Nathan Chappell, 00:48:29.250 --> 00:48:31.869 thank you so much for your time today. Absolute 00:48:31.869 --> 00:48:34.329 pleasure. We'll catch up soon. Yeah, definitely. 00:48:34.750 --> 00:48:35.929 Let's do it again soon. Thanks.