1 00:00:00,240 --> 00:00:07,649 [Auto-generated transcript. Edits may have been applied for clarity.] Uh, this lecture today is part of the MSC in Translational Health Sciences that we are running here 2 00:00:07,650 --> 00:00:13,890 at the University of Oxford on specific themes of technological innovation and digital health. 3 00:00:14,370 --> 00:00:20,490 Uh, so Nicole is an associate professor in business and society at the National College of Ireland. 4 00:00:20,910 --> 00:00:23,790 Uh, we're really happy she's going to talk to us about power, 5 00:00:23,790 --> 00:00:35,040 politics and the fight for justice in the age of a and really hot topic that's been discussing, uh, some of these issues earlier in the week. 6 00:00:35,370 --> 00:00:40,620 Uh, Nicole is really widely established. She works with researchers across Europe as well. 7 00:00:40,620 --> 00:00:47,760 So you bring that perspective to Nicole and people to hearing about your work. 8 00:00:48,360 --> 00:00:52,379 So over to you. We have a battle now. The next going to take about an hour. 9 00:00:52,380 --> 00:00:55,500 And then we have five minutes for questions and discussion. 10 00:00:56,250 --> 00:01:03,809 Perfect. Well, thank you so much for having me. Um, I'm delighted to be there because I think this is a really important topic to talk about. 11 00:01:03,810 --> 00:01:07,049 And, uh, I, I lecture myself. 12 00:01:07,050 --> 00:01:12,420 So I do tell this stuff to my own students to, uh, but also in other capacities, 13 00:01:12,420 --> 00:01:18,990 whether it's publishing or policy papers, I try to bring, uh, the same message across in different forms. 14 00:01:19,830 --> 00:01:28,319 Um, so my goal today is really to tell a story, um, and the story is, um, about, um, 15 00:01:28,320 --> 00:01:34,410 digital health because digital health is so much more just, um, technologies or tools or devices, 16 00:01:34,860 --> 00:01:40,350 uh, because it's just easy to pick off, uh, a theme, so to speak, and think that's digital health, 17 00:01:40,800 --> 00:01:44,850 but it is really a story about power and politics and the fight for justice. 18 00:01:45,330 --> 00:01:49,830 And because I want to tell a story, I have divided it into different parts. 19 00:01:50,130 --> 00:01:56,040 So the first one is like, how did we get here? Because obviously, you know, um, everything has a trajectory. 20 00:01:56,040 --> 00:02:03,359 And the trajectory is important in many ways because, um, unless major interventions happen, 21 00:02:03,360 --> 00:02:08,700 like, especially when it comes to power and politics, things tend to go into similar direction. 22 00:02:09,090 --> 00:02:12,090 And that's called often in the literature part dependency. 23 00:02:12,660 --> 00:02:16,410 So um, and then obviously I have three different parts here to it. 24 00:02:16,410 --> 00:02:20,250 And uh, your team had asked me to present on three of my papers. 25 00:02:20,700 --> 00:02:24,780 So the first one is called choreographing for Value in Digital Health. 26 00:02:25,230 --> 00:02:31,379 Uh, the second one is a potent problem indeed, where I talk a little bit more about AI because this is so relevant at the moment. 27 00:02:31,380 --> 00:02:38,940 And then the last one is healthcare at a crossroads. And then, of course, wouldn't be a good story without a contemplative pause. 28 00:02:39,300 --> 00:02:48,240 And then a question, where do we go from here? Um, so I suppose all good stories start with, uh, once upon a time, uh, tagline, of course. 29 00:02:48,240 --> 00:02:52,200 And I'll do that too. Okay, so as I said, every story has it gone. 30 00:02:52,410 --> 00:02:58,030 Once upon a time moment. So, um, this scenario happened in Ireland and Ireland. 31 00:02:58,030 --> 00:03:07,290 That's important here for that kind of context because, um, Ireland is obviously lauded as the tech capital or the Silicon Valley of Europe. 32 00:03:07,770 --> 00:03:14,190 Um, so once upon a time in a research centre, uh, in Ireland, they, uh, that was in around 2015. 33 00:03:14,190 --> 00:03:24,659 Here you can see of, of a press release, um, there was, uh, an industry and academia collaboration, um, where basically people from industry, 34 00:03:24,660 --> 00:03:32,399 the state healthcare providers and academics came together to try to make sense of digital health and try to, 35 00:03:32,400 --> 00:03:37,950 you know, bring it into the market and also to up, uh, public health and public value, 36 00:03:37,950 --> 00:03:42,330 really, uh, for your convenience, have obviously highlighted a few key words here. 37 00:03:42,690 --> 00:03:51,030 Um, it was talking about a proactive model connecting stakeholders, acute care patient at the centre, empowering patients, um, 38 00:03:51,030 --> 00:03:57,330 having good interventions in the care pathway here and then scale, um, personalised health care, 39 00:03:57,570 --> 00:04:01,200 you know, patient experience and reducing costs, which is all very important. 40 00:04:01,200 --> 00:04:09,420 And of course, we know, um, that Europe and many other countries are on a big crash course in terms of ageing populations, 41 00:04:09,720 --> 00:04:16,590 um, health care constraints. So, of course, you know, given that kind of team health care or public health was, uh, divisive. 42 00:04:17,130 --> 00:04:20,820 So that's what, uh, I work in this research centre. 43 00:04:20,820 --> 00:04:23,790 Of course. That's why I kind of use this as a as a theme. 44 00:04:24,360 --> 00:04:30,750 So the vibe here was that we are coming together to bring, uh, digital technologies into public health. 45 00:04:31,230 --> 00:04:36,090 And of course, uh, the the promises are always very, um, enticing. 46 00:04:36,090 --> 00:04:41,760 Here, now, sorry, this is not working again for me to go to the next slide. 47 00:04:41,770 --> 00:04:46,680 Oh, here. Lovely. So the promises, of course, of digital health are very enticing. 48 00:04:47,010 --> 00:04:53,640 Um, and they still are to this day. And I even to this day, I do believe, truly, that digital health has enormous, 49 00:04:53,910 --> 00:04:59,820 enormous capabilities to make public health better when it comes to likes of disease intelligence wanted. 50 00:05:00,270 --> 00:05:02,480 Clinical decision making, patient centred care. 51 00:05:02,490 --> 00:05:10,580 And indeed we see many, many benefits of public health and digital technologies already in public health right now, and even patient empowerment, 52 00:05:10,590 --> 00:05:14,040 having a doctor in your pocket and as a mom, as a matter of fact, 53 00:05:14,040 --> 00:05:19,860 like many of us have health apps and they use AI even know to kind of hack their health. 54 00:05:20,520 --> 00:05:26,579 Um, of course there's big promises of operational efficiencies, better quality of care at speed and at scale, 55 00:05:26,580 --> 00:05:33,840 cost savings, etc. and of course that people get better and faster if they work in health care too. 56 00:05:34,680 --> 00:05:45,210 So, um, moving on here, you can see here, um, with the best intentions in the research centre, we started researching and advising and executing. 57 00:05:45,750 --> 00:05:53,280 Um, and you can see a picture of me here, and I be very happy. Uh, I that's because, like, we we had problems along the way, too. 58 00:05:53,610 --> 00:06:00,840 And, uh, this is a picture of my, uh, lovely co author, Susie Geiger there that you will find on many of my papers here on our paper, 59 00:06:01,080 --> 00:06:05,520 combined papers, and of course, many other researchers. It's never a team. 60 00:06:05,700 --> 00:06:10,109 It's it's always a team effort. So I worked at the time in a particular unit. 61 00:06:10,110 --> 00:06:15,090 So this gives me probably a little bit of a tainted view as well because there was different units at the research centre. 62 00:06:15,570 --> 00:06:20,840 But the questions we were asked all the time is what is the best business model, 63 00:06:20,850 --> 00:06:25,169 you know, if if by the companies, uh, and what's a good market entry strategy? 64 00:06:25,170 --> 00:06:29,430 How do I sell to the national health care system or, you know, 65 00:06:29,640 --> 00:06:36,510 in Ireland and also and so in Europe and in fact, what how do I sell my testing anywhere? 66 00:06:36,840 --> 00:06:41,669 Because our experiences with some of the technology companies were that they did develop their 67 00:06:41,670 --> 00:06:47,460 products in isolation of care pathways or revenue models or even what's going on in public health. 68 00:06:48,180 --> 00:06:57,059 Um, that kind of gave us, in some ways problems, because we went into the research centre with the promise of improving public health. 69 00:06:57,060 --> 00:07:00,150 So we were asking questions, what what about connecting stakeholders? 70 00:07:00,150 --> 00:07:04,440 What about patients and care pathways? What about improving health care? 71 00:07:04,440 --> 00:07:10,079 And remember public health, of course. Um, and which was very interesting because like, um, 72 00:07:10,080 --> 00:07:17,520 fast forward a little bit of time and the research centre field and we also written about this afterwards in 2022, 73 00:07:17,820 --> 00:07:22,620 um, in a paper called Leaning in or Falling Over What it takes to make a market. 74 00:07:23,070 --> 00:07:26,390 Um, so this was actually quite complicated because, like, 75 00:07:26,400 --> 00:07:32,430 there was a big mismatch between what they wanted and what we wanted to give them or what we wanted as academics. 76 00:07:32,850 --> 00:07:41,909 And we ended up working and in a kind of state of liminality where we didn't know what what was happening and if you were coming and going. 77 00:07:41,910 --> 00:07:49,559 But it was very chaotic at the time to, um, in saying our experience as a research centre wasn't unique, really, 78 00:07:49,560 --> 00:07:53,730 because all the companies were trying to get into health care at that at that stage, 79 00:07:53,730 --> 00:07:58,440 big and small, uh, and there were struggling and there was many reasons for that. 80 00:07:58,440 --> 00:08:03,509 It could have been that they as I said, they didn't understand, uh, power structures in health care. 81 00:08:03,510 --> 00:08:09,780 They didn't understand national health care systems, pathways, revenue models or revenue pathways. 82 00:08:10,290 --> 00:08:14,999 Uh, but they wanted to sell their thing there, take things, so to speak, of a solution. 83 00:08:15,000 --> 00:08:21,360 Uh, it cetera without understanding necessarily, how to do a good value proposition that address public health. 84 00:08:21,950 --> 00:08:28,229 And as I said, the research centre ended up failing. So, um, moving on here. 85 00:08:28,230 --> 00:08:32,760 So from 2014 onwards, we saw some quite concerning shifts. 86 00:08:33,270 --> 00:08:40,889 Um, we tracked them over to time to. So it was called e-health, mHealth, connected health or digital health at the time. 87 00:08:40,890 --> 00:08:44,190 That's a term that we seem to have settled now in, in the literature. 88 00:08:44,670 --> 00:08:50,250 But it was truly a wave, a hype and a voyage. And why I'm saying that is we did. 89 00:08:50,460 --> 00:08:53,820 You can see here a paper on the right hand side. It was called this hype. 90 00:08:53,820 --> 00:09:02,130 Uh, creative irreversibility. And you, you need a lot of promises and investments to build a market. 91 00:09:02,520 --> 00:09:11,040 And at that time in the mid 20 tens, these investments often came from the state who supported that, 92 00:09:11,040 --> 00:09:16,380 but also from the tech companies, of course, too, because they wanted to find a market for their technologies, 93 00:09:16,920 --> 00:09:21,239 but because it was so difficult to get into the public health market, 94 00:09:21,240 --> 00:09:27,809 many of them defaulted into the direct to consumer market, which at that time was growing quite heavily, too. 95 00:09:27,810 --> 00:09:33,389 You might remember two times of, you know, your first fitness tracker or your first health app, 96 00:09:33,390 --> 00:09:40,680 etc. and of course, the the health care system also lauded patient empowerment a lot at the time, 97 00:09:40,680 --> 00:09:43,620 saying, well, it's not just the doctor that needs to tell you what to do, 98 00:09:43,620 --> 00:09:47,850 but you also need to take care of your own health care and your own responsibility. 99 00:09:48,540 --> 00:09:54,479 So at the same time, we also saw an increasing data application and digitalisation of health care, 100 00:09:54,480 --> 00:09:59,790 which is not a problem in itself, but it came alongside an increase in growth of surveillance. 101 00:09:59,820 --> 00:10:07,830 And data capitalism. Shoshana Zuboff has written a fabulous book about this, um, which is well worth reading for anybody who's interested. 102 00:10:08,250 --> 00:10:18,120 Uh, but the idea is that basically data like that, the data is stripped of people and to pervasive surveillance, 103 00:10:18,120 --> 00:10:22,379 you know, you like you have lots of data trails in your daily life behind. 104 00:10:22,380 --> 00:10:26,820 And we know that. And but this money, this data is converted into assets. 105 00:10:27,180 --> 00:10:33,600 And by assets I mean like, you know, things you can sell on the market and have real current, uh, earnings from that. 106 00:10:34,080 --> 00:10:41,460 And of course, when you think about now the advertising market like it's Google and Meta and many other advertisers, 107 00:10:41,880 --> 00:10:49,830 they basically absorb your data and then sell it on the market for, you know, for it's called the market for future behaviours. 108 00:10:50,070 --> 00:10:56,550 To any companies that want to know what makes you tick. But they basically tell all, take all your data and sell it. 109 00:10:57,120 --> 00:11:01,650 Um, and of course, and then the advertisers use it for, you know, targeting stuff back at you. 110 00:11:02,220 --> 00:11:08,880 Um, so we track that in multiple industries, for example, in the consumer genomics industry where you, 111 00:11:08,910 --> 00:11:13,620 you know, the 23 to me and ancestry etc., that they have a multi-sided platform, 112 00:11:13,620 --> 00:11:22,259 you pay for the service, but you also kind of get your data in the background if you agree to terms and conditions or at times even to research. 113 00:11:22,260 --> 00:11:29,010 And the consumer doesn't know what research is, so to speak, if they think they're helping medical science, um, at the same. 114 00:11:29,010 --> 00:11:31,890 And also we looked at that in the context of mental health apps. 115 00:11:32,400 --> 00:11:41,610 Uh, we saw also very worrying, um, developments in, in the increase of power of big tech and colleagues like, uh, 116 00:11:41,610 --> 00:11:48,660 Tom or Sharon and colleagues, they have talked, they have written about this quite extensively, and I call this a sphere transgression. 117 00:11:48,990 --> 00:11:54,660 So when big tech companies started to get into health care, into public services, 118 00:11:54,660 --> 00:12:02,250 they're kind of getting into spaces where tech companies probably should not have as much input and power. 119 00:12:02,700 --> 00:12:10,560 Um, and what also became very clear that there was a very, very clear move away from public health promises and promises. 120 00:12:11,040 --> 00:12:13,739 Um, and that is, of course, a worrying development. 121 00:12:13,740 --> 00:12:23,370 So, um, at that stage, of course, along came Covid and that was in around 20, 2020, as you most of you will remember. 122 00:12:23,910 --> 00:12:27,540 And then, of course, there were so many questions that came along with covet. 123 00:12:27,540 --> 00:12:34,320 And of course, as academics, we take these questions and we we use them as an opportunity to do more research. 124 00:12:34,920 --> 00:12:38,129 So our questions were at the time, you know, when Covid hit. 125 00:12:38,130 --> 00:12:43,709 Well, what what was a public health crisis of this scale due to the adoption and scale of digital health? 126 00:12:43,710 --> 00:12:49,800 And many of you might remember the, uh, the Covid tracking systems or booking system. 127 00:12:49,800 --> 00:12:58,230 So everything all of a sudden got almost digitalised overnight because we had no other choice because of the contact, low contact kind of environment. 128 00:12:58,830 --> 00:13:06,239 Um, and of course, at the same time, you know, because the state was or the many of the European states were a bit out of their depth. 129 00:13:06,240 --> 00:13:09,480 They also asked, uh, tech companies to come on board to help. 130 00:13:10,260 --> 00:13:16,140 Um, so we saw public, uh, public private partnerships flourishing. 131 00:13:16,530 --> 00:13:19,769 But we asked ourselves, will this mark a return to public health? 132 00:13:19,770 --> 00:13:27,959 Because we saw this big crisis and we asked ourselves, well, will this be a good a bad thing for public health at last, as their power shifts? 133 00:13:27,960 --> 00:13:36,450 And then obviously, as we know from the contact tracing within Europe, on the back end of that, the European health data space was developed. 134 00:13:36,960 --> 00:13:42,120 Um, or or not, of course, because these are always questions that are remain to be seen. 135 00:13:42,120 --> 00:13:46,170 And that's what I meant by my point in the slide that it is a journey too. 136 00:13:46,770 --> 00:13:50,730 So we ended up taking that into the deeper as we do, of course. 137 00:13:51,150 --> 00:13:55,800 So digging deeper here meant, oh, sorry, let me go back. 138 00:13:56,190 --> 00:14:05,009 Um, the, the in previous papers we had pointed to finger quite well at big tech, you know, 139 00:14:05,010 --> 00:14:10,380 and that these that sphere of transgressions and they were having more and more powers. 140 00:14:10,770 --> 00:14:17,129 But we also asked ourselves at that point, well, you know, what role does the EU or state policies really, 141 00:14:17,130 --> 00:14:21,480 uh, have in choreographing these public private entanglements? 142 00:14:21,990 --> 00:14:27,270 And they are actually entanglements, uh, we call them actually techno tangles, 143 00:14:27,270 --> 00:14:32,670 which are push and pause or dance, as we call them, a dance between the EU and Big Tech. 144 00:14:33,180 --> 00:14:37,950 Because it is a back and forth, it's never a clearer, you know, just one contract. 145 00:14:37,950 --> 00:14:41,280 It is a back and forth because they in some ways they need each other. 146 00:14:41,370 --> 00:14:45,719 Like in many economies, public value and market value is intrinsically entangled. 147 00:14:45,720 --> 00:14:50,760 And this is one way of seeing, um, digital health to us as an entanglement. 148 00:14:51,240 --> 00:14:57,240 And the other question that we had, of course, is how does can this dance orientated towards public health? 149 00:14:57,240 --> 00:15:04,770 Because this is something that we wanted to. Keep in mind from day one, in telling the story to that public, how it is here, what matters. 150 00:15:05,670 --> 00:15:10,499 So what's the problem here? Well, you know, there was many issues at play. 151 00:15:10,500 --> 00:15:17,579 For example, um, the entrepreneurial state, um, many of the we know that from America anyway. 152 00:15:17,580 --> 00:15:22,350 But many of the European countries also they play entrepreneurial state. 153 00:15:22,650 --> 00:15:29,010 And Ireland is particularly guilty in that respect with our foreign divestment, uh, foreign direct investment strategies. 154 00:15:29,340 --> 00:15:36,780 And it's also very clever in many ways to of course. Um, but by playing Entrepreneurial State, um, you are encouraging these techno tangos, 155 00:15:36,780 --> 00:15:41,240 but they always come with consequences to benefits, promises and perils. 156 00:15:41,300 --> 00:15:49,410 Yeah. Of course. Um, and Mariana mazzucato has written quite a lot about the entrepreneurial state and quite critically too. 157 00:15:50,040 --> 00:15:59,460 Uh, so we have also a big data driven health economy by now that we cannot talk about healthcare anymore without talking about, uh, data. 158 00:16:00,300 --> 00:16:07,860 Um, so of course, uh, and there was also state driven data sales already in the UK happening at this time. 159 00:16:08,460 --> 00:16:16,620 And of course, there is a quite unstoppable blending between public health and new digital, uh, concepts and tools that, you know, as I said, 160 00:16:16,620 --> 00:16:19,979 you can't talk anymore about healthcare but are talking about digital health, 161 00:16:19,980 --> 00:16:24,330 and that will be coming to that point in time where you can't talk about it anymore without talking about AI. 162 00:16:25,050 --> 00:16:32,730 So and the looming question here is, well, what public value is actually created in and through those health tangles? 163 00:16:33,120 --> 00:16:35,520 And that is a difficult question to ask. 164 00:16:35,820 --> 00:16:45,150 So what we did is we went through 30 years of documentation from the EU and did a discourse analysis pretty much to 165 00:16:45,210 --> 00:16:54,270 see about that kind of notion of public value and how it is directed and shaped and shifted and how this can happen. 166 00:16:54,690 --> 00:17:01,889 And the question was also, of course, what does the European health data space achieve here in terms of the public health and public values now? 167 00:17:01,890 --> 00:17:04,230 But of course, what is public value anyway? 168 00:17:04,230 --> 00:17:10,920 Um, it's the question how public sector institutions build a further their democratically established schools. 169 00:17:11,280 --> 00:17:13,349 And again, when it comes to healthcare, 170 00:17:13,350 --> 00:17:22,650 we always have to remember that healthcare is very special because health care is a public good and a human right as defined by the W.H.O. 171 00:17:23,070 --> 00:17:29,070 That means the state has a social contract to its citizens to provide health care. 172 00:17:29,520 --> 00:17:32,850 And of course, how does it improve the still lives of citizens. 173 00:17:33,300 --> 00:17:39,810 And it's also a way of measuring progress towards achievement, of widely accepted societal goals. 174 00:17:40,290 --> 00:17:44,579 And of course, the state always has to be proactive and value creating in, 175 00:17:44,580 --> 00:17:51,180 in, in creating these kind of, um, improvements and to further social goals. 176 00:17:51,510 --> 00:17:58,530 And our question was, well, how does the state do that over the course of 30 years and all their policies to do with digital health, 177 00:17:59,430 --> 00:18:11,100 how are their mission oriented policies now? The grim realities are in many ways, that's when it comes to sorry, what's that question? 178 00:18:12,960 --> 00:18:18,000 Okay. Uh, the grim realities are in many technologies. 179 00:18:18,030 --> 00:18:23,160 Even if you hold your own mobile phone here in many technologies. 180 00:18:23,190 --> 00:18:25,290 Even, as I said to your smartphone, 181 00:18:25,290 --> 00:18:33,929 the fundamental technological tools and infrastructures that went into that were actually developed by public investments, uh, 182 00:18:33,930 --> 00:18:37,589 by the state investing in these technologies, however, you know, 183 00:18:37,590 --> 00:18:44,970 it is the apples and the Samsungs of this world who really monetise these technologies, and it's informed by the same thing. 184 00:18:45,420 --> 00:18:49,230 There's many public investments that have gone into fundamental R&D. 185 00:18:49,650 --> 00:18:56,309 Um, but they have been turned into assets by, by pharmaceutical companies, for example, 186 00:18:56,310 --> 00:19:03,720 and Covid alone, nine new Covid billionaires where basically, you know, um, born. 187 00:19:04,110 --> 00:19:12,780 But many of the as I said to you, the vaccine technologies or the vaccine research was actually done by public investments and universities. 188 00:19:13,380 --> 00:19:17,610 So at the same time, we have also a lot of infrastructural monopolies, 189 00:19:18,000 --> 00:19:23,790 even think about your own kind of, uh, in healthcare, I think about the power of big tech, 190 00:19:23,790 --> 00:19:31,050 all the emails to the devices, to infrastructures, the cloud space, all these infrastructures, 191 00:19:31,320 --> 00:19:36,960 uh, often, uh, you know, owned and held by a handful of big tech companies. 192 00:19:37,320 --> 00:19:41,070 And at the same time, we have a what's called a lot of data lockouts. 193 00:19:41,070 --> 00:19:49,379 So they are taking charge of the data, uh, in the absence of knowing who actually owns the data. 194 00:19:49,380 --> 00:19:55,320 And, but they're also excluding the public from this data and from what's, what is supposed to be an information commons. 195 00:19:55,320 --> 00:20:00,090 And what I mean by that is that data in general could be a common good. 196 00:20:00,330 --> 00:20:05,070 It could be owned by everyone because you own your own data too. 197 00:20:05,520 --> 00:20:10,349 Um, but yet you have no idea where it goes half the time and what happens with it. 198 00:20:10,350 --> 00:20:17,219 So we lose, we lose sight of it. Um, and of course, there is pervasive and increasing surveillance. 199 00:20:17,220 --> 00:20:20,700 Um, and also what's called a data enclave. 200 00:20:20,700 --> 00:20:23,099 And Cambridge has written about this, 201 00:20:23,100 --> 00:20:31,679 about data enclaves to these are all protected spaces that big tech built to keep this data safe and for themselves, 202 00:20:31,680 --> 00:20:38,430 and that other people can have access to them. And of course, during Covid, we had a large new wave of Covid. 203 00:20:38,430 --> 00:20:43,470 Techno tangles and new norms were established here that are very difficult to overturn. 204 00:20:44,280 --> 00:20:48,089 Now, the European health data space is interesting here because yes, 205 00:20:48,090 --> 00:20:57,060 there is definitely much more medical research and innovation because of that, because now we are either at a position where data is shared. 206 00:20:57,390 --> 00:21:06,480 That was before. Nor I can also say enclaves, you know, within different countries, but now it's more shared, which is which is a good thing. 207 00:21:06,870 --> 00:21:11,790 Um, so there is much more opportunity for public health research and innovation too. 208 00:21:11,790 --> 00:21:16,260 But at the same time, we have also a big European data lake now. 209 00:21:16,770 --> 00:21:27,540 And when you think back now of the story I've told you so far about sphere transgressions and and technology companies accessing data, 210 00:21:27,960 --> 00:21:31,710 you know, quite freely, then you can see that this is a problem. 211 00:21:31,710 --> 00:21:37,620 And of course, many of when I look back at the documentation, many of this is cloaked in snazzy business language, 212 00:21:37,620 --> 00:21:41,040 like old data for innovation and opportunities to enter new markets. 213 00:21:41,490 --> 00:21:49,170 And there is nothing new here because looking at, uh, the, the documents over 30 years ago, quite clearly that this language is very pervasive, 214 00:21:49,170 --> 00:21:55,139 much more than the language of public health or public value, of course, that it's very, very complicated. 215 00:21:55,140 --> 00:21:59,400 Questions, too, about who directs these policy conversations and how is this done? 216 00:21:59,970 --> 00:22:04,470 So colleagues here, um, and I'm show you a really quickly, I'm going to pop out of this for a second. 217 00:22:04,950 --> 00:22:05,970 Colleagues here. 218 00:22:06,390 --> 00:22:17,100 Um, but this, this great, uh, Dutch repository here, um, which I have written quite eloquently here by presenting big tax law lobby playbook. 219 00:22:17,640 --> 00:22:19,410 And they have done it for different countries. 220 00:22:19,410 --> 00:22:25,170 So these are all case studies here about the narratives and, you know, echo chambers and stuff like that. 221 00:22:25,170 --> 00:22:31,200 So it's very, very good. These are all in the presentation. So you can click into it again at your convenience to to look at it more. 222 00:22:31,200 --> 00:22:34,080 I'll share a presentation afterwards for to share with the audience of course. 223 00:22:34,830 --> 00:22:40,260 Um, but this is a great repository which shows you how they do their lobbying. 224 00:22:40,260 --> 00:22:43,080 And the dangers, of course, are committed to. 225 00:22:43,980 --> 00:22:51,600 So, um, and of course, there's loads of interesting models here too, and of course that the pharma, sorry, 226 00:22:51,600 --> 00:22:58,050 the big tech playbook and the lobbying influences has influenced quite heavily also the European health data space. 227 00:22:58,560 --> 00:23:07,200 Now the interesting models that um, that came because of that was the the primary and secondary use of the data that flows within Europe. 228 00:23:07,200 --> 00:23:12,300 Now. Now the primary use is everything medical really any. 229 00:23:12,510 --> 00:23:16,410 To do with health and medical stuff, and you can't really opt out of this. 230 00:23:16,430 --> 00:23:20,800 This data flows either way. But there's also a secondary use of data. 231 00:23:20,810 --> 00:23:24,520 And this is what what is a lot more difficult to understand. 232 00:23:24,530 --> 00:23:30,710 It is the stuff for data for innovation and markets and marketing that is a lot harder to grasp, 233 00:23:31,310 --> 00:23:35,000 uh, for an ordinary person, you know, for the ordinary Joes. 234 00:23:35,000 --> 00:23:38,210 So that clicks agreed to all cookies that they see, right. 235 00:23:38,630 --> 00:23:42,810 Um, because people are not as informed us as, you know, as we should. 236 00:23:43,120 --> 00:23:50,270 They should be. And then this also this questions about opting in or opting out of the secondary use of data. 237 00:23:50,600 --> 00:23:56,990 So the initial question was should people be automatically opted in and then they have to opt out, 238 00:23:57,440 --> 00:24:05,839 or should they be, you know, um, just, you know, should you opt out the another opportunity or option? 239 00:24:05,840 --> 00:24:12,469 So what they have done is now so basically should people have to specifically opt in or specifically opt out? 240 00:24:12,470 --> 00:24:19,100 And what they have decided now is that everybody is kind of opted in, but it has to be an opt out option. 241 00:24:19,700 --> 00:24:29,210 But they left it to every country now Ireland, you know, the Netherlands, Germany, France to to implement that for themselves. 242 00:24:29,630 --> 00:24:32,420 So how that looks like now is up to every country. 243 00:24:32,420 --> 00:24:39,290 So there will be fragmentation, but many people will not understand that the ordinary Joe open not understand. 244 00:24:39,290 --> 00:24:45,830 If they are sitting in the doctors and they are asked, would you like to participate in research with your data? 245 00:24:46,640 --> 00:24:50,959 Many people are very altruistic and go like, oh well, I can do something good here. 246 00:24:50,960 --> 00:24:58,550 So yes, they don't understand that necessarily. What that means that they should be opting out or what what any of the implications of that are. 247 00:24:59,090 --> 00:25:04,309 And the question is also when it comes to secondary data, you know, 248 00:25:04,310 --> 00:25:11,840 the premise is it can be used by by, you know, invite the by the market, by tech companies, by, 249 00:25:11,840 --> 00:25:15,440 you know, anybody who, who has an interest in health care, 250 00:25:15,560 --> 00:25:22,280 as long as it has a benefit to individuals and society, as long as it develops, basically delivers public value. 251 00:25:22,700 --> 00:25:25,880 But the question is, how is that actually policed and accounted for? 252 00:25:26,240 --> 00:25:32,000 Who who checks up on this and what does that actually mean. And that's these are the the complicated question. 253 00:25:32,000 --> 00:25:35,510 Many of the European health data space now. 254 00:25:36,680 --> 00:25:44,250 Moving on here. So what? Solutions are possible. So is it possible to redirect value flows back to the public? 255 00:25:44,270 --> 00:25:47,980 They cannot be public private partnership or technical. 256 00:25:48,000 --> 00:25:52,700 Technical tangles should not be based on unbalanced or parasitic relationships. 257 00:25:53,060 --> 00:25:53,780 They should be. 258 00:25:53,780 --> 00:26:00,950 There could be profit sharing clauses, conditionalities in public contracts or digital tax that could be redirected back to public health. 259 00:26:01,520 --> 00:26:07,040 But it needs political will also by the state to put this into contracts. 260 00:26:07,520 --> 00:26:12,560 And at the moment, you know, as you can see from this kind of act or chapter, 261 00:26:12,920 --> 00:26:19,010 that the state has a lot of responsibility also to enable the companies to do what they do. 262 00:26:19,550 --> 00:26:26,540 There could be also a public value would be beyond financial returns or to do an exercise of public value mapping. 263 00:26:27,080 --> 00:26:31,010 My colleagues here have a data solidarity in the University of Vienna, 264 00:26:31,010 --> 00:26:38,450 have a data solidarity mapping tool where you can click into it and see how much public value does. 265 00:26:38,570 --> 00:26:40,850 What does the data even generate? 266 00:26:40,880 --> 00:26:48,740 Some of them don't even generate any public data at public public value, so it's not probably even worth, you know, worrying about this too much. 267 00:26:48,740 --> 00:26:52,100 But some data really does and that needs to be looked at. 268 00:26:52,730 --> 00:26:59,870 You could also have a common space approach to how the data to have to make sure that these data enclaves or that these, 269 00:26:59,870 --> 00:27:03,140 these protected spaces, these monopolies are disbanded. 270 00:27:03,530 --> 00:27:08,810 So you could have, for example, data sharing pools or data courts of public data trusts. 271 00:27:09,200 --> 00:27:16,520 But the point is they shouldn't be all owned and managed and and held together by the big tech companies. 272 00:27:17,300 --> 00:27:22,160 So of course then along came AI or generative AI. 273 00:27:22,160 --> 00:27:27,290 And I'm sure you all remember this when the gen AI models started scaling. 274 00:27:27,290 --> 00:27:34,950 And of course, now we are in a position where they have crept into every tool, into every tool bar, into every application. 275 00:27:34,970 --> 00:27:42,490 AI is somewhere there. And of course, with that, there's even more questions of, you know, it is a truly transformative technology. 276 00:27:42,500 --> 00:27:45,830 So will it change digital health and how will it change due to health? 277 00:27:46,400 --> 00:27:52,400 And of course, how it will this transformation mean a return to public value at last? 278 00:27:52,400 --> 00:27:56,570 And finally, and of course, is there any issues emerging here? 279 00:27:56,570 --> 00:28:01,400 And then of course we we dug deeper and I put this little robot here. 280 00:28:01,430 --> 00:28:04,909 Of course he is. Of course, robots are not the same as AI. We know that. 281 00:28:04,910 --> 00:28:08,750 But you might want to remember him or her or this this bot here. 282 00:28:09,530 --> 00:28:16,130 Okay, so a potent problem indeed. So this is the second act or the second chapter I want to talk about. 283 00:28:16,730 --> 00:28:22,850 And of course, you've seen it even now in the media, AI is the promised land. 284 00:28:23,090 --> 00:28:28,520 Of course we wonder why we had this heard this one before. We have heard this exactly ten years ago with Digital Health, 285 00:28:28,880 --> 00:28:39,920 where we had all these promises which are needed to basically and investments to build a market and to build or to let these technologies flow into, 286 00:28:40,430 --> 00:28:49,399 you know, the applications or real life. And the research questions here, where, um, how does generative AI, um, impact health care? 287 00:28:49,400 --> 00:28:54,980 What data does this bring to bring along? And how how can these be addressed really? 288 00:28:55,280 --> 00:28:57,230 And I'll explain data justice in a minute. 289 00:28:57,260 --> 00:29:05,480 Of course, it is a transformative technology, and it brings along so many complex and layered ethical hurdles that affect health, wealth and power. 290 00:29:06,080 --> 00:29:11,420 Um, so data justice in this most simplest form is a fairness by which people are made 291 00:29:11,420 --> 00:29:16,040 visible or presented and treated as a result of their production of digital data. 292 00:29:16,550 --> 00:29:19,710 And we all produce digital data. 293 00:29:19,730 --> 00:29:26,540 You cannot be any more existing in the world without being online in some shape or form, even older generations now. 294 00:29:26,750 --> 00:29:33,110 But you think about by 2030 all public services are online basically. 295 00:29:33,110 --> 00:29:35,960 So it doesn't matter. You have to participate in this ecosystem. 296 00:29:36,650 --> 00:29:43,820 What is meant by visibility of by Taylor is this access to representation to information 297 00:29:43,860 --> 00:29:49,879 privacy by by being represent that it means how do you engage with the technology. 298 00:29:49,880 --> 00:29:54,770 Is the other benefits shared and is there autonomy in technology choices? 299 00:29:54,770 --> 00:30:01,330 And even what I said a minute ago, like if everything is online, it is probably not even so many, uh, 300 00:30:01,610 --> 00:30:08,060 choices anymore because we are already catapulted in an in an element of choice lessness because everything is online. 301 00:30:08,690 --> 00:30:13,070 And then of course, um, the last bit is how are you being treated? 302 00:30:13,080 --> 00:30:18,770 Is that non-discrimination? Can you challenge bias and can you prevent discrimination? 303 00:30:19,130 --> 00:30:27,800 Now, Taylor extended their framework in 2025 on data justice to include preserving and strengthening, uh, 304 00:30:27,800 --> 00:30:34,160 public infrastructure structures and courts because they realise that this is so much bigger than just the individual. 305 00:30:34,700 --> 00:30:39,320 Um. A small ecosystem, but it is truly, uh. 306 00:30:39,950 --> 00:30:42,920 How technology flows into public goods and infrastructures. 307 00:30:43,370 --> 00:30:50,479 And also they also extend it to inclusiveness, to contestability and accountability and to global responsibility. 308 00:30:50,480 --> 00:30:54,020 And we will come to that, um, as we go along further in our story here. 309 00:30:54,950 --> 00:30:58,040 Now, um, so what is the problem here? 310 00:30:58,070 --> 00:31:02,180 Well, truly, it is a pick and mix. Um, there is a problem. 311 00:31:02,180 --> 00:31:04,100 And we've seen that this before. 312 00:31:04,100 --> 00:31:12,919 As a team of unchecked power, Big Tech has been establishing power and making money from data as asset for over 20 years. 313 00:31:12,920 --> 00:31:15,559 So they have a very strong to track to be here. 314 00:31:15,560 --> 00:31:22,430 And of course, they have a lot of political influence, as you saw earlier with the with the big tech labour and lobbying. 315 00:31:22,910 --> 00:31:30,590 There's also this, uh, big problem of equity that these processes of and practices of data extraction 316 00:31:30,590 --> 00:31:35,720 are so mundane and pervasive and hidden that we don't even understand these. 317 00:31:36,260 --> 00:31:40,850 And I'm sure if I ask you in the room, do you read terms and conditions? 318 00:31:41,270 --> 00:31:45,559 People will shake their head. Do you accept cookies, people? 319 00:31:45,560 --> 00:31:49,610 But not because it is about convenience, about clicking things away? 320 00:31:50,150 --> 00:31:58,820 Um, and of course we don't even know how many how much data we shared, whether it's like this topic type, the stuff, uh, 321 00:31:58,820 --> 00:32:06,979 you know, even our, you know, the stuff we say, our voice, our movements, GPS tracking, and we even have now neuro tech. 322 00:32:06,980 --> 00:32:09,440 That's where it reads your brainwaves. 323 00:32:09,770 --> 00:32:16,069 And it's other things too, like, you know, uh, there's also sensors and, you know, think about stuff like smart cities, 324 00:32:16,070 --> 00:32:22,010 many things are doing without permission because they're just in bins or in roads, sensors, etc. 325 00:32:22,550 --> 00:32:28,730 And when you think about the more ambient technology becomes, the more invisible it becomes, 326 00:32:28,730 --> 00:32:35,510 because it is so seamless and frictionless and integrated in our lives that we don't see it anymore. 327 00:32:35,510 --> 00:32:39,680 And this also includes your AI technologies, new AI agents. 328 00:32:40,280 --> 00:32:48,559 Um, of course, there's a problem of access. You know, that maybe we have access to all these technologies, but not everybody doors. 329 00:32:48,560 --> 00:32:52,040 Some people don't have phones. They don't have internet, they don't have literacy. 330 00:32:52,460 --> 00:32:56,480 They don't have to live in low and middle income countries. It's totally different. 331 00:32:56,900 --> 00:33:01,640 And there is this inequality and institutional and structural discrimination. 332 00:33:02,060 --> 00:33:05,660 And of course, resources as we know are not distributed fairly in the world. 333 00:33:05,660 --> 00:33:09,290 And the most marginalised are by default most more disadvantaged. 334 00:33:09,290 --> 00:33:12,290 Uh, and they will be more disadvantaged even further by AI. 335 00:33:13,040 --> 00:33:16,399 So there's also this problem of identity. 336 00:33:16,400 --> 00:33:26,150 And you might want to remember my bot here. Um, so every time I would have said to you, if you picture a robot or I as a person, 337 00:33:26,510 --> 00:33:30,980 you would have given me probably that description of the male white robot, right? 338 00:33:31,460 --> 00:33:36,680 Because this is, you know, there are so many biases naturally inbuilt in the AI system, 339 00:33:36,680 --> 00:33:42,860 whether it's like gender biases, there might be racial biases, political language biases. 340 00:33:42,860 --> 00:33:45,650 Because many of these technologies come from the West. 341 00:33:46,160 --> 00:33:52,640 Um, there's a fundamental lack of representation that is harmful categorisations of self from communities. 342 00:33:53,150 --> 00:33:57,620 Um, there is an the ratio of identity, and it is not inclusive by design. 343 00:33:58,010 --> 00:34:03,530 And often it is starting literally from the technology design all the way down to the implementation. 344 00:34:03,950 --> 00:34:08,299 Um, and of course there is a problem with participation too, that, you know, 345 00:34:08,300 --> 00:34:14,600 there's many dehumanising and oppressive features, like even when you think about other AI technologies, for example, 346 00:34:14,600 --> 00:34:22,370 like, um, you know, in hiring or in H or many bots nowadays, scan your CV, 347 00:34:22,400 --> 00:34:29,210 there's many you don't even get to be selected because the AI has already rejected your application from the get go. 348 00:34:29,720 --> 00:34:38,150 There's also lots of choice lessness and dark patterns, as I said, for there's not always a choice there that algorithmic decisions are being made. 349 00:34:38,840 --> 00:34:43,520 Even if you don't participate actively in AI because they are running around in the background, 350 00:34:43,820 --> 00:34:49,880 you know, because there's so much data around you and of you at any moment in time, 351 00:34:50,180 --> 00:34:59,030 because it is coming from different data sources, but also over time, because we are online for such a long amount of time now in our lives, 352 00:34:59,270 --> 00:35:06,110 from being born or from even before your being born, even, you know, your mother had, you know, 353 00:35:06,110 --> 00:35:11,000 if you a mother takes now a pregnancy test or has a period tracking up, 354 00:35:11,330 --> 00:35:16,040 then the system knows you're pregnant before your parents do or before your doctor knows. 355 00:35:16,370 --> 00:35:23,800 So again, it is. It's all such a long amount of time that the companies have been able to to build themselves up. 356 00:35:23,810 --> 00:35:32,060 Um, um, you know, a vision of you, um, again, and there's a lot of choice listeners too, because, like you don't you cannot always opt out of things. 357 00:35:32,330 --> 00:35:35,960 And there's a lot of dark patterns, too, and many papers have been written about that. 358 00:35:36,020 --> 00:35:43,280 And you probably know that two that you've heard recently about the dark patterns in social media and the addictive features. 359 00:35:43,790 --> 00:35:51,380 We are now already at the point where we have AI, empires, and tech companies owned these AI empires. 360 00:35:51,560 --> 00:35:57,350 So again, the AI is just an extension of the powers that they already have. 361 00:35:57,380 --> 00:36:06,610 It supplements that ecosystem. So it's very, very difficult as an individual or it's just me or you to challenge the status quo of power, 362 00:36:06,620 --> 00:36:10,729 because as I said, it is so pervasive and mundane that we often don't see it. 363 00:36:10,730 --> 00:36:17,360 And because we feel as a very small person in a very big ecosystem that we don't even have any powers over. 364 00:36:17,660 --> 00:36:22,240 So fair participation isn't always there. And then, of course, the knowledge too. 365 00:36:22,250 --> 00:36:27,950 I mean, the AI systems were implemented with them move fast and break things kind of approach. 366 00:36:28,370 --> 00:36:35,269 Um, even if there are times have killed people and I'm sure you have heard that in teenagers in America, 367 00:36:35,270 --> 00:36:40,339 but also in Europe, have used these systems to vent their mental health issues. 368 00:36:40,340 --> 00:36:45,200 And we're kind of almost nurture towards suicide, which did happen at the end. 369 00:36:45,650 --> 00:36:48,650 So, you know, it is quite serious. 370 00:36:48,650 --> 00:36:53,030 And of course, there's black boxes and algorithms as trade secrets, secrets. 371 00:36:53,030 --> 00:36:57,139 And big tech is unwilling to share knowledge and to foster this intercultural learning 372 00:36:57,140 --> 00:37:00,620 because they want to protect their business and revenue models and their trade secrets, 373 00:37:00,620 --> 00:37:05,320 of course. Now there's also grim realities about data justice here. 374 00:37:05,330 --> 00:37:08,700 We know that Pandora's box cannot be closed. 375 00:37:08,720 --> 00:37:11,900 I cannot be invented at this point. Right? 376 00:37:12,560 --> 00:37:15,680 Um, there's also an AI arms race. 377 00:37:16,100 --> 00:37:20,630 So, for example, the US has 40 models, India has 18. 378 00:37:20,820 --> 00:37:29,420 Um, I think 15 or 17 models. Uh, China has 25 or 27 and U.S. four models, Foundation models. 379 00:37:29,420 --> 00:37:31,670 So it is a two AI arms race at the moment. 380 00:37:32,150 --> 00:37:39,980 But because of that arms race, they have chosen to gloss over the risks and harms and issues that are kind of clearly emerging. 381 00:37:40,700 --> 00:37:45,110 Um, and there's also a next level of surveillance and data capitalism. 382 00:37:45,140 --> 00:37:54,469 So before when you think about it, but before when you search something, you put a search term in or you asked your search engine something. 383 00:37:54,470 --> 00:38:03,650 But now we have a lot more interaction with AI, which means that, you know, we we have we have a back and forward conversation. 384 00:38:03,650 --> 00:38:07,220 So it's not just there's a lot more data coming from you. 385 00:38:07,490 --> 00:38:17,420 And also now when you think about it's a genetic AI or uh, AI agents, they have not just access to one application on your computer, but to many. 386 00:38:17,900 --> 00:38:23,299 Um, so again, there is a more high level surveillance opportunity there to look at absolutely 387 00:38:23,300 --> 00:38:27,890 everything that you have on your desktop or on your computer or on entire hospitals, 388 00:38:27,890 --> 00:38:32,270 computers, let's say. And yet there's no such thing as a fair value in return. 389 00:38:32,750 --> 00:38:36,220 There's also, of course, a lack of transparency and accountability. 390 00:38:36,230 --> 00:38:40,280 But yet you have these AI empires now, which huge amounts of money. 391 00:38:40,880 --> 00:38:43,910 And there's also a lack and representational justice. 392 00:38:43,910 --> 00:38:51,440 Like there is such a growing evidence of harms, misuse databases, data security and privacy issues. 393 00:38:51,450 --> 00:38:57,590 Um, so that quest of obtaining fairness and dignity and autonomy is getting harder to do. 394 00:38:57,920 --> 00:39:03,710 And even stuff like, you see, as I said, all the AI tools have crept into your tool box. 395 00:39:03,950 --> 00:39:06,290 So whether or not you like them, they're just there now. 396 00:39:06,290 --> 00:39:14,860 And Google has announced yesterday you probably saw today their AI that that their normal Google search will be all a lot more AI infused. 397 00:39:14,870 --> 00:39:21,610 So you can't even do a normal Google search anymore. Of course, there's huge, um, intellectual property issues. 398 00:39:21,620 --> 00:39:31,159 So think about when I hoovered in the learning and training data sets over all the books, all the paintings, 399 00:39:31,160 --> 00:39:39,260 all the artists stuff, all the poems, and hoovered all this up without any any regard to intellectual property issues. 400 00:39:39,260 --> 00:39:50,060 So the people like artists and writers and also newspapers and etc., academics, books, you know, it was completely cast aside. 401 00:39:50,300 --> 00:39:52,250 There was no fairness, autonomy there. 402 00:39:52,730 --> 00:39:59,180 And of course, there is this pervasive presence of entrepreneur capitalism tool, but a lack of Covid with a lack of global regulation. 403 00:39:59,780 --> 00:40:05,839 So there's of course more problems here. There is flaws in restorative justice. 404 00:40:05,840 --> 00:40:10,550 But these these are the processes and mechanisms that you need for reconciliation and repair. 405 00:40:11,660 --> 00:40:16,190 Well, I mean, we have loads of high level ethical concepts and guidance here. 406 00:40:16,190 --> 00:40:26,000 But to be honest, the big tech companies, they it's meaningless and toothless and isolated and useless if it isn't regulated properly. 407 00:40:26,420 --> 00:40:33,170 Um, so ethical guidance is great, but it is not really binding. 408 00:40:33,650 --> 00:40:42,049 Of course you can have also individual. Like, you know, you can, for example, report about output abuses or wrong doings of to the I directly. 409 00:40:42,050 --> 00:40:46,910 You've probably seen these that you can answer back saying that's not correct, or you can give thumbs up and thumbs down. 410 00:40:47,540 --> 00:40:52,010 Um, or you can launch a lawsuit. But this is complicated in its own way. 411 00:40:52,190 --> 00:40:56,959 To whom exactly? To the tech company, to the developers, to the I. 412 00:40:56,960 --> 00:41:01,070 Who do you launch a lawsuit against? And what's an AI crime anyway? 413 00:41:01,400 --> 00:41:06,920 So because AI is fairly new, these are all these questions that people just don't have the answers for. 414 00:41:06,920 --> 00:41:14,270 And that's why we had put the robot open jail, so to speak. There's this AI incident database that I just want to show you to quickly. 415 00:41:15,540 --> 00:41:22,530 So there is this, um, I incident database here, this global one where you can report, 416 00:41:23,130 --> 00:41:28,680 uh, I incidents on an ongoing basis to, to map them and tracked them. 417 00:41:29,100 --> 00:41:36,030 Uh, there's also, um, for example, organisations and researchers. 418 00:41:36,210 --> 00:41:44,680 So I've talked only last week to the Amsterdam Institute for Global Health and Development and the lovely Gloria, 419 00:41:44,700 --> 00:41:48,060 uh, umbrella and her, her team and a supervisor. 420 00:41:48,600 --> 00:41:56,100 So they have, for example, developed here an AI maturity, AI governments and maturity ranking website, 421 00:41:56,100 --> 00:42:01,110 where they looked at a world map here, where they looked at the maturity levels. 422 00:42:01,110 --> 00:42:07,019 Here was frameworks and what jurisdictions and what levels they're at, you know, 423 00:42:07,020 --> 00:42:14,520 and you can see here many countries do have some regulation, but many of them also don't or it's incomplete. 424 00:42:15,180 --> 00:42:23,070 They have also developed here um, uh, a matrix here, a global health AI governance matrix here, 425 00:42:23,400 --> 00:42:28,110 where they looked at different tiers and how binding they are. 426 00:42:28,110 --> 00:42:31,620 And when you go, it's quite interesting here because when you go a bit further down here, 427 00:42:31,620 --> 00:42:36,899 you can see here there's hardly any binding international treaties or supernatural things. 428 00:42:36,900 --> 00:42:44,520 There's locally binding laws, but not really anything on a global level that is binding or supernatural. 429 00:42:44,850 --> 00:42:54,480 Then this sort of research is really, really important because this gives us an an idea of the state of play where we are at at the moment. 430 00:42:54,660 --> 00:42:59,580 So there is jurisdictional regulation, but nothing supernatural or nothing global. 431 00:42:59,580 --> 00:43:06,209 And that's a problem because what big tech or tech companies do is they look for pockets of opportunities, 432 00:43:06,210 --> 00:43:10,440 of course, uh, because that is basically their business model and their market. 433 00:43:10,920 --> 00:43:15,870 Um, and there's also more obviously grim realities here that way, because. 434 00:43:16,800 --> 00:43:23,920 You know what needs to be changed in order for fairness and and human flourishing to happen in theory. 435 00:43:23,920 --> 00:43:27,280 And look again at all the promises of AI that we see. 436 00:43:27,640 --> 00:43:32,379 There is clear benefits, including the potential to change people's dignity and well-being, 437 00:43:32,380 --> 00:43:40,000 and also to make to enhance public life and public health, and to bring everybody on an even keel when it comes to education. 438 00:43:40,420 --> 00:43:48,130 But, you know, there is so many harms at the same time and so many injustices that at the moment it is an ethical minefield. 439 00:43:48,490 --> 00:43:53,680 So that flourishing is not not facilitated by the big tech companies. 440 00:43:53,680 --> 00:43:58,810 And regulation, as we could see, is still inconsistent or lagging behind. 441 00:43:59,260 --> 00:44:04,810 So what solutions are there? Of course, transformative technologies do need transformative solutions. 442 00:44:05,440 --> 00:44:10,600 So for example, um, peace, emancipation, eliminating the costs of injustice. 443 00:44:11,080 --> 00:44:15,630 Um, you know, what we need is what I showed you the website earlier. 444 00:44:15,640 --> 00:44:18,790 We need something legally binding and global agreements. 445 00:44:19,150 --> 00:44:25,660 If we can do intellectual property regulations like the WTO, the World Trade Organisation has it. 446 00:44:26,020 --> 00:44:34,480 Why can't we have legally binding regulations when it comes to AI and protecting people from injustices like that? 447 00:44:34,960 --> 00:44:41,950 And also, we have to remember that modern conflicts are a push and pull between, uh, status and power. 448 00:44:41,960 --> 00:44:49,900 So I think the state can learn, for example, from what happens today when it comes to stuff like, 449 00:44:50,170 --> 00:44:58,030 you know, the social media errors that they have made. The state can also learn, for example, where to apply political will better. 450 00:44:58,300 --> 00:45:04,150 And citizens also need to learn from conflict, for example by engaging more in social movements. 451 00:45:04,510 --> 00:45:11,819 And because we are often so busy on our own phone, we forget that there is, you know, we should bind together and be more solidarity. 452 00:45:11,820 --> 00:45:15,460 Stick with others. Uh, in terms of human rights approaches. 453 00:45:16,150 --> 00:45:20,020 This panel principles, there should be glass boxes, not black boxes. 454 00:45:20,350 --> 00:45:26,919 We should enforce active participation. We should allow big tech in in lobbies for regulation. 455 00:45:26,920 --> 00:45:32,499 And for example, when you remember and I hope you do remember, but eventually, over the years, uh, 456 00:45:32,500 --> 00:45:38,050 big tobacco like the the companies that used to do the cigarettes, 457 00:45:38,290 --> 00:45:45,850 they got kicked off the the political decision making table, um, because of the harms caused. 458 00:45:46,150 --> 00:45:51,400 And why can't Big Tech be kicked off the lobby when it comes to regulation making? 459 00:45:51,730 --> 00:45:55,380 And also states need to question their, uh, 460 00:45:55,390 --> 00:46:03,100 entrepreneurial state approach and that innovation and commerce friendly stands at all costs and I mean at all costs. 461 00:46:03,490 --> 00:46:08,050 I think it's important to facilitate innovation, commerce, but not at all costs. 462 00:46:08,680 --> 00:46:12,280 And, of course, you know, when it comes to empowering agents and people, 463 00:46:12,280 --> 00:46:18,100 it's like looking at the the lived experience, advocacy work, citizens, juries or other AI reporting. 464 00:46:18,700 --> 00:46:23,080 It needs to be set up at which to challenge power differentials here. 465 00:46:23,440 --> 00:46:26,989 And. The state need all states ont. 466 00:46:26,990 --> 00:46:34,340 EU needs to listen much more to what people have to say and not just the, you know, the tech companies now. 467 00:46:35,210 --> 00:46:42,290 And of course then more AI. And we know that because in recent years like it has scaled significantly. 468 00:46:42,650 --> 00:46:47,000 And of course there might be also what about real data justice issues. 469 00:46:47,000 --> 00:46:53,329 And as we know, there's more regulations coming like the AI act has kicked in. 470 00:46:53,330 --> 00:46:55,700 Now the Digital Services Act, 471 00:46:55,700 --> 00:47:06,260 which is about looking at control and choice for consumers and platforms and decrease illegal harmful content and digital markets Act is about, 472 00:47:06,650 --> 00:47:12,890 you know, identifying these exactly these gatekeeper platforms, the big tech platforms and holding them responsible. 473 00:47:13,220 --> 00:47:20,990 And there's also such thing as the Digital Fairness Act coming, or in the making or in the discussions at the moment, uh, which looks at, 474 00:47:20,990 --> 00:47:28,700 you know, the marketing side of art, dark patterns, influencer marketing, addictive design and the unfair personalisation practices. 475 00:47:29,120 --> 00:47:34,129 Um, you know, anything to exploit consumer vulnerabilities for marketing purposes. 476 00:47:34,130 --> 00:47:39,830 And the big question is, are changes coming? So of course, again, we talk deeper here as we do. 477 00:47:40,370 --> 00:47:46,250 And we looked at health care at the crossroads here. And yes, of course the promises are still going very strong. 478 00:47:46,730 --> 00:47:53,450 But we look this time more into that. What I mean, what I mentioned before are these entanglements, uh, 479 00:47:53,450 --> 00:47:58,370 between health care markets and regulation and who and what drives these entanglements, 480 00:47:58,760 --> 00:48:05,180 and how can we also, again, more or more towards justice rather than data capitalism? 481 00:48:05,870 --> 00:48:12,379 So the theoretical lens that we used here, and I don't want to go too much into the sociology side of it, 482 00:48:12,380 --> 00:48:17,450 but that that these are basically socio material assemblages, 483 00:48:17,450 --> 00:48:23,959 which means there are an entangled mess of networks that are human, non-human, material, immaterial, 484 00:48:23,960 --> 00:48:29,570 and physical and textual things, and they all interact with each other on an ongoing basis. 485 00:48:29,870 --> 00:48:35,779 So it's people, materials, actors, body spaces, but it's all about control. 486 00:48:35,780 --> 00:48:43,040 And who directs these conversations and how and what's what becomes contested as a consequence. 487 00:48:43,460 --> 00:48:49,100 So the problem here is that we know that China is so entangled in health care and markets and regulation, 488 00:48:49,100 --> 00:48:57,700 and that the big tech companies are transgressing spheres, and the technology often acts as an overlay or underlay here. 489 00:48:57,780 --> 00:49:06,259 What I mean by that is that technology is absolutely everywhere. But it's not just technology, it is power plays, and it is markets, 490 00:49:06,260 --> 00:49:12,770 and it's also politics and political struggles are all imprinted into these technologies. 491 00:49:13,190 --> 00:49:16,820 So these are also on the move and these are creating ripple effects. 492 00:49:17,240 --> 00:49:20,270 Um, it's we call this an ontology of movement. 493 00:49:20,270 --> 00:49:25,040 And there's two opposing concerns here. So one of them is data justice. 494 00:49:25,730 --> 00:49:28,520 What I mentioned before and what came from Taylor, 495 00:49:29,030 --> 00:49:36,290 which were the digital technologies and algorithms and data and health and medicine are already putting the most marginalised in society, 496 00:49:36,290 --> 00:49:44,690 um, further disadvantage. But there's also this big issue of data capitalism, which is the modus operandi of the tech companies, where data, 497 00:49:44,900 --> 00:49:52,100 including your personal health data, is transformed into assets and monetised with lock outs and all these inequalities. 498 00:49:52,580 --> 00:50:00,860 Now, we've had all these before, but AI is adding so much new layers of actors and interests and powers and revenue here, huge amounts. 499 00:50:01,430 --> 00:50:06,530 So the thing is, and I said that from the very start, digital technology is not just technology. 500 00:50:07,340 --> 00:50:13,729 It AI is in flux, it is moving, and it can be directed and exploited by actors with powers and resources. 501 00:50:13,730 --> 00:50:20,720 And when you look back again at the previous two acts that I presented as part of the story, 502 00:50:21,080 --> 00:50:27,350 you can see already the key actors here and the powers and resources, um, it is persistently provisional. 503 00:50:27,350 --> 00:50:34,969 It's invisible, unstable, and it's being constantly configured, can be configured, which is a good and bad thing because it's so hard to pin down. 504 00:50:34,970 --> 00:50:39,800 What, because it's also provisional means it can also be reconfigured towards the good. 505 00:50:40,550 --> 00:50:45,290 So these ripple effects, they affect, you know, health care delivery regulation, 506 00:50:45,290 --> 00:50:50,930 tech markets, they destabilise other actors and move them in the process. 507 00:50:51,260 --> 00:50:54,739 So even by talking, uh, to health care providers, they said, well, 508 00:50:54,740 --> 00:50:59,600 we've lost basically control over essential infrastructures and we don't even know anymore. 509 00:50:59,600 --> 00:51:06,620 We're losing all this knowledge because when it comes to essential updates of technologies or training and implementation, 510 00:51:06,620 --> 00:51:10,250 we don't have this knowledge anymore. They do, the tech companies do. 511 00:51:11,090 --> 00:51:15,649 And of course, you know, citizens often complain saying, I can't even participate in that because, 512 00:51:15,650 --> 00:51:19,850 you know, I don't understand it anymore and I can opt out and I can say no. 513 00:51:20,270 --> 00:51:24,660 So again, over the last 20 years, when you think about that push and pull of. 514 00:51:24,710 --> 00:51:26,780 To two different directions. This movement, 515 00:51:27,260 --> 00:51:36,500 it has been moved and often also with the state's help towards surveillance and data capitalism and data enclaves and data ecosystems and monopolies. 516 00:51:36,800 --> 00:51:42,550 And the EU has a responsibility here to, uh, to to let that happen. 517 00:51:42,560 --> 00:51:47,510 And of course, there has been attempts over the years to pull towards the data towards the side. 518 00:51:47,900 --> 00:51:57,650 And I showed you some, some resources here, like the as I said, the mapping out to incidence databases or public policy papers. 519 00:51:58,250 --> 00:52:03,500 You know, there's loads of efforts also from people like me and my colleagues, this academic activism. 520 00:52:04,190 --> 00:52:13,010 Now the grim realities is that it is AI is the big tech companies are still encroaching public health even further. 521 00:52:13,550 --> 00:52:17,180 And because they make so many promises, 522 00:52:17,180 --> 00:52:27,470 it's easy to get swept away by these promises and to allow them to get more foothold in, in, in, in public health. 523 00:52:27,800 --> 00:52:35,000 And of course, AI is an unstable actor. There is more mounting evidence of biases, hallucinations and a lack of value. 524 00:52:35,450 --> 00:52:44,299 And, you know, I'm sure you've all heard of the dead internet of of the AI collapse because the the more AI generated content there is, 525 00:52:44,300 --> 00:52:47,870 the more issues they are with the accuracy of their outputs. 526 00:52:48,380 --> 00:52:51,530 Um, and of course there is issues with consent and privacy. 527 00:52:51,860 --> 00:52:59,150 And of course, people feel very much undermined in terms of visibility and autonomy, and there's still a big lack of oversight and accountability. 528 00:52:59,630 --> 00:53:06,290 People are often very choice too. And I know all this legislation in place, like or in the making, like the Digital Fairness Act, 529 00:53:06,290 --> 00:53:10,160 that where the EU was at least aware of it, other geographies are not. 530 00:53:10,700 --> 00:53:13,360 There is often a lack of options not turn ative steers, 531 00:53:13,430 --> 00:53:19,640 TS and CS cookies consent opting out digital literacy issues where we are feeding varied choices as consumers. 532 00:53:20,360 --> 00:53:22,300 Um, and of course, as I said, 533 00:53:22,310 --> 00:53:30,890 it is a huge data capitalism boost because there so much more data in and through the AI and AI cements Big Tech's powers even more. 534 00:53:31,370 --> 00:53:37,639 There has been a lovely book written by Ivan on ABC by Bruno Mucha, who do talks about world builders, 535 00:53:37,640 --> 00:53:42,560 and his argument is that, uh, the world is built anymore by states. 536 00:53:42,830 --> 00:53:47,810 Uh, it is built by big tech companies who have more power and more money than individual states. 537 00:53:48,140 --> 00:53:53,930 And the regulation, as yet is yet not there to rebalance these power symmetries significantly. 538 00:53:53,930 --> 00:54:04,430 So what are solutions here? Of course, we need to empower civic society more and have more literacy skills building and core creation, 539 00:54:05,510 --> 00:54:12,709 less surveillance, less of a like I can't even say less automation because we are definitely getting more automated, 540 00:54:12,710 --> 00:54:17,840 but at least to facilitate that understanding and have also clear and options to opt out, 541 00:54:18,530 --> 00:54:24,500 we also need to move towards dynamic, more dynamic regulation and more precaution of regulation. 542 00:54:24,510 --> 00:54:25,670 Sometimes that works well. 543 00:54:26,120 --> 00:54:34,670 For example, we have a new digital omnibus, which is good, and that I will talk about this a little bit more in a minute, but. 544 00:54:35,840 --> 00:54:39,320 You've probably saw the recent notify fire ups that. 545 00:54:39,350 --> 00:54:42,230 You know that came out which was very harmful obviously. 546 00:54:42,620 --> 00:54:54,379 But in the the new omnibus they have made that immediately, um, uh, illegal to do this, uh, notify our UPS or these new eyes, which is really good. 547 00:54:54,380 --> 00:55:00,890 So that dynamic element is good, but we still always go much slower than what what the tech companies can release. 548 00:55:01,370 --> 00:55:07,339 And also, I think we need to decide what is off the menu when it comes to AI and health care. 549 00:55:07,340 --> 00:55:10,940 For example, what about AI generated viruses? 550 00:55:10,940 --> 00:55:14,059 What about palliative care? What about assisted suicide? 551 00:55:14,060 --> 00:55:17,510 What about full automation of care? Where do we set? 552 00:55:17,870 --> 00:55:22,519 Where do we have precautionary regulation? Where we say we do not allow dogs the same way? 553 00:55:22,520 --> 00:55:28,000 We don't allow, for example, um, you know, human embryo cloning or other things, right? 554 00:55:28,010 --> 00:55:32,059 Or, you know, um, some experimentation on humans, etc. 555 00:55:32,060 --> 00:55:35,300 So there has to be really clear guidelines about what's off the menu. 556 00:55:35,960 --> 00:55:43,580 We also need to and I've said that before and all the other solutions we need to push for public health, uh, public value accountability. 557 00:55:44,020 --> 00:55:47,930 I should not disproportionately provide a big tech. 558 00:55:48,620 --> 00:55:52,100 Um, so a contemplative pause at this point. 559 00:55:52,820 --> 00:55:57,680 Where are we now? Um, so I, I put a picture up there. 560 00:55:57,680 --> 00:56:02,450 I was almost upset that it wasn't good news first and then bad news because I like talking about bad news. 561 00:56:02,690 --> 00:56:06,980 But I'm actually glad it's this way, because I think I want to focus on the good news, too. 562 00:56:07,490 --> 00:56:10,610 So we know. And the bad news site. My favourite, of course. 563 00:56:11,120 --> 00:56:15,110 Um, there is a continuous and aggressive push by Big Tech, 564 00:56:15,590 --> 00:56:22,790 and we see that even now the investments that are made, there's also still very worrying sphere transgressions. 565 00:56:22,790 --> 00:56:31,600 And I'm sure you've all heard about that recent case of Palantir getting more access to own anonymized NHS patient data. 566 00:56:32,240 --> 00:56:37,040 This is definitely not what you want, especially since Palantir's very clear about marketing data. 567 00:56:37,550 --> 00:56:41,480 We are also now in a point where we have a lot of shadow health care systems. 568 00:56:41,780 --> 00:56:47,389 And what I mean by that is like, we have a lot of, uh, we don't just have a public healthcare system anymore. 569 00:56:47,390 --> 00:56:54,080 We have people hacking their mental health on ChatGPT or character I or Google Gemini. 570 00:56:54,080 --> 00:57:00,889 We have a shadow health care system that's completely unregulated. We have a lot of job losses and environmental costs of AI. 571 00:57:00,890 --> 00:57:06,320 And I want to talk about health care here in a broader terms as well, because health care is not just what happens in the doctors. 572 00:57:06,590 --> 00:57:11,719 If you lose your job. Your job is your health. It's got to be effective. Think about the environmental cost of AI. 573 00:57:11,720 --> 00:57:14,660 This is all going to affect our health and wealth in the long term too. 574 00:57:15,200 --> 00:57:23,179 We have AGI scare us, and I hope you've all heard of notebook, which was, uh, this social media platform for AI agents. 575 00:57:23,180 --> 00:57:28,190 And the first one thing why it became famous is because so many people are. 576 00:57:28,190 --> 00:57:31,640 Well, uh, but but why don't we overwrite and kill humans? 577 00:57:32,180 --> 00:57:33,890 So this is, of course, very worrying. 578 00:57:34,370 --> 00:57:41,930 We have also the digital omnibus proposal and what that is, it's it's a move by the EU to, I don't want to say essentially deregulate, 579 00:57:42,320 --> 00:57:48,860 but it is pretty much about supporting businesses, reducing compliance and costs and supporting innovation. 580 00:57:48,860 --> 00:57:53,750 Again, all the stuff we've heard before and also kicking the timelines now to August 2028. 581 00:57:54,440 --> 00:58:00,259 And of course we have stuff like Trump politics, which is all very bad news, uh, because it doesn't have full global regulation. 582 00:58:00,260 --> 00:58:06,589 But now here's the good news. The critical voices are getting louder and you can see more and more. 583 00:58:06,590 --> 00:58:11,719 And I hope you all are getting the critical voices, too. And it's not just me sitting in some kind of filter bubble, 584 00:58:11,720 --> 00:58:18,320 which is all very possible to with my searches and stuff that I do, but the critical voices I feel are getting louder. 585 00:58:18,710 --> 00:58:24,050 There's also good regulation in the EU kicking in, but don't forget EU is the gold standard here. 586 00:58:24,050 --> 00:58:26,600 Um, not every place in the world is the same. 587 00:58:27,170 --> 00:58:36,320 We have also moved towards theorising the digital terms of health, and I'm also writing a paper at the moment with colleagues, uh, 588 00:58:36,830 --> 00:58:44,989 a global network of colleagues for The Lancet, where they're looking at the digital as a structural condition that shapes chiefs health. 589 00:58:44,990 --> 00:58:51,500 And we are looking at all the things that I've discussed here, the business models, the political side of the political norms, 590 00:58:51,500 --> 00:58:56,570 the infrastructures, the personal practices of us, important determinants of health. 591 00:58:57,140 --> 00:59:05,990 There's also big litigations. Now, you saw the landmark ruling about Meta and Google, about the social media, um, you know, the harms it has done. 592 00:59:06,470 --> 00:59:12,320 And these landmark rulings are really important now because whereas public health 593 00:59:12,320 --> 00:59:18,470 or technology somehow should not be based on litigation based approaches, 594 00:59:18,860 --> 00:59:24,500 but it is at least a start to show that big tech companies are doing the harms. 595 00:59:24,950 --> 00:59:34,170 There's also local solutions. You might not have seen it, but Palantir got kicked off, um, in the UK also for, uh, you know, tech. 596 00:59:34,320 --> 00:59:41,370 Analysis for refugee accommodation, and they source their own, um, open source software system for that. 597 00:59:41,640 --> 00:59:45,660 They keep Palantir of that because they are aware of these sphere transgressions. 598 00:59:46,080 --> 00:59:51,969 There's also national solutions. Now, we saw Australia doing the social media ban and given the harm, 599 00:59:51,970 --> 01:00:00,270 and starts to take the eye of many of the digital technologies doing for certain cohorts in particular, like younger people. 600 01:00:00,270 --> 01:00:03,510 It's really important to also extend this maybe to AI. 601 01:00:04,140 --> 01:00:09,930 And as you could see here, from what I showed you from the different in my different from the different acts in the story, 602 01:00:09,930 --> 01:00:14,460 there is coordinated efforts there for global regulation and digital governance of health. 603 01:00:15,300 --> 01:00:19,500 Now, finally, a last slide. Concluding remarks. 604 01:00:20,220 --> 01:00:29,910 So what do we need? I tried to, in the story, try to the main message home that we need public value in digital health, 605 01:00:30,540 --> 01:00:36,110 not just buzzwords like innovation and markets and all this. 606 01:00:36,120 --> 01:00:38,070 We need public value and public health. 607 01:00:38,550 --> 01:00:44,879 But what we have is dependencies on big taken power asymmetries and we that are very difficult to dislodge at this point in time. 608 01:00:44,880 --> 01:00:51,090 But they are not ever irreversible. And it is also adding new complications and layers. 609 01:00:51,090 --> 01:01:00,030 Here. What needs to happen is coordinated and systematic, practical and dynamic contestation of power, profits and politics and health care. 610 01:01:00,660 --> 01:01:05,100 We all need to fight and regulate for fairness, justice, equity and democracy. 611 01:01:05,490 --> 01:01:08,400 And of course, I can't leave without giving an inspirational quote here. 612 01:01:09,030 --> 01:01:18,059 So Shoshana Zuboff has said, if the digital future is to be our home, then be it is or step or we that who need to make it. 613 01:01:18,060 --> 01:01:25,170 So we will need to decide who decides. And this is the fight for our for our future and future. 614 01:01:25,590 --> 01:01:32,070 Um, even though we are often faced with choice lessness and with a big daunting task, we have to remember that. 615 01:01:33,460 --> 01:01:39,670 All angles, whether it's student scorn, you know, get an education and going into health, 616 01:01:39,820 --> 01:01:46,180 into the marketplace with a different insight or citizens organising themselves for, 617 01:01:46,210 --> 01:01:57,190 you know, advocacy and citizen movements or the state regulating more dynamically and more precautionary or tech companies taking on ethical guidance. 618 01:01:57,580 --> 01:02:01,230 It is it is all possible, but it is a fight. 619 01:02:01,240 --> 01:02:09,250 We have to remember that. So I want to I want to leave it here at that at the talk and open the floor up for questions. 620 01:02:09,250 --> 01:02:09,730 Of course.