Of PDIS. And her work focuses on open civic knowledge, digital democracy, future of education, public service, collaborative policy, and rulemaking. And, I'll let them introduce their project for today. But before we start, did you have any questions on we are the group? Oh, and now that Nathan is here, can I just request that Nathan, would you mind recording the session so they can share with their team after?

Oh, excellent. Thank you. So and I forgot to mention the general structure presentation for as long as you need. But towards the end, I'll keep an eye on the chat session, and we'll have some time for discussion or however you would like to run it. But it's quite flexible. If you wanna pull this group for questions or just do a presentation, however you whatever is helpful for you.

Yeah. It's it's it's great. Great. It's really nice to be here with such a really diverse interdisciplinary community, if we can name it as a community. Regarding the questions, you can interrupt anytime, or we can leave it to the end or we can have other conversations towards the end, but just feel free to interrupt with us. I guess since we are not too many, so we have the chance to kind of having questions while we run through. Yeah. Fang, you can go ahead. I will support you.

Yeah. If I may jump here briefly just to manage the expectations. So what we are presenting here is, as Feng has mentioned, is narrow bit of overall project and also in a very a very surface level. But then later on in the conversation, we will give links to, the materials where, then we can the the where we it's possible to find more details. Yeah. Sorry for interruption. Go ahead.

Yeah. Maybe it's good to slightly mention why we have this eco versus eco. So if you look into, like, the past history of human or in planet itself, maybe there maybe the current dominant more verbal perspective is the hierarchical layered parliament where men, not even women, one part of the human being is on the top of the pyramid that actually could be can be seen dominant in the cultures, in in the governance models, in the society, and in the economic systems we we work through. But the urgency of the climate change, along with other societal urgencies that we are experiencing these days, actually, requires a mind shift. And then what we are trying to also highlight that, that the the the from a holistic ecological mind shift is just not new. It's not that something we have invented. There's, existing cultures, currents and past, cultures, that's actually supporting that different more holistic mind shift, actually, the the genes of that kind of perception and understanding of the society. On top of that, while we are talking about augmented intelligence, we also have the ability to use to that use or leverage advancements in AI, in data, or technology in general, but not for the sake of technology, but within the context. And even there are evidences of that kind of systemic entanglement in initial sciences, in physics. The the studies in quantum physics is also is a good domain to explore and understand, which requires a different perspective. So either we look into culture or society, we look to the technology or science itself. We we we we have evidences and ingredients to actually support that kind of mind shift and which basically looking into interdependencies at micro level, but also the how that emerges with in in mac macro level. So that's maybe a brief assumptions, a background that guide has guided our is guiding our research or and and all of it.

Yeah. Maybe before moving into a graph description of the framework itself, we are all aware that when we deal or interact with climates or environment in general, when we are trying to solve the problem itself, there are different approaches from mitigation to adaptation and as well as how we can have resource or society level or resource level resilience. And on in within whole solution space, we see that governance is also very, very essential. So the the framework we are thinking or we are we we are looking into can be applied in any of these domains. Although in this particular conversation, we are focusing on how we can bring and expand the governance with how we can actually attribute agency to to to things or beings that was not part part of conversation or was not part of decision making process. So in the end, what we are kind of want to share here is is a framework. As we know, the framework has a a theoretical paradigm, a set of tools, a vision, and an objective. So what we are looking into here is just looking into basic building blocks of these, more tangible components that may help, to incorporate existing, developments in AI and in in in data science in in general. So when we talk about the AI data science or similar technical augmentation, we are very consciously aware of two facts. One of them is that we are they really try to be sensitive that we are we don't want to have technology driven, techno driven solutions, but as well as looking into cultural value in their aspects of that domain. So from technical sides, there are emerging positive, encouraging developments that can be adopted, such as a good survey could be climate change AI with they have very extensive survey of possible machine learning tools that can be used in in in the domain. Basically, you could use any solutions, any technological solutions, any, algorithmic solution in the in, within it. But it's not only technology. It's also good positive signs. There are, for instance, platforms or initiatives that, put in efforts, thoughts into how we can use, the AI or or, any relevant domains for social goods or for ecology. Even big organizations like Google's, they also kind of in in in the process. So these are all good signs. But what we are having here, the challenge is how can we integrate communities, neighborhoods, society in in in larger extent. When we pick one or the other solutions, how are we addressing other problems other than the climate? What how how are we dealing with the, the social aspect and economical aspects, and how we are actually integrating explicitly ethical principles in in in the process. So so these are the kind of surrounding element that's we we think is worth to mention. As as of technical aspects, here we just list a nonexhaustive selections of opportunities. So but at the same time, it's they are not completely random because if you want to be able to have, societal aspects, individuals' participation, deliberations process, and also extend this with a perspective of more than human elements, then that large space of solutions can be narrowed down. For instance, if we when we involve in data collection, use of data, helping behavioral change, it's important to pay attention to privacy of individuals. So, in that case, using technologies like differential privacy or different learning methodologies where you don't have to look into a raw data, combine patterns in a more anonymous way, like using federated learning techniques, naturally will come out more when when we look to that perspective. And when we want to expand the data streams, which could increase awareness and understanding of other beings around us, then IOTs, the edge AI computing, or technologies around that has actually support will be supporting it. And then when we look into let's say, like, in the example we are looking into how we can give voice to a bee, a voice to river with existing capacity of sensors and understanding the process, is it possible to capture the dynamics of how a tree interacts with its environment and model it even in a hypothetical virtual environment using reinforcement learning, for instance, using distributed AI techniques where you can go beyond one tree and be able to see how those three three entities are interacting with each other and then do and explore possibilities. And so that's that that kind of ex examples basically give us a perspective of what choices we may be looking into. I guess there's someone is asking. Yeah. If if they're asking if we can stop for a question. Yeah. If it's something really or that's that is good to have the have and also, it's it's it's important to to understand it. Let's go ahead for the question. Otherwise, we can leave it to the end. Okay. Then, we can, leave the discussions. So within we are we just we have mentioned a kind a a space how we can bring in different approaches and different solution spaces. It at the same time, we are the way we are looking into how we combine machines and humans is important to understand where where decision process is taking place, how machines and the humans will be interacting. So an example would be to what extent humans are more autonomous versus to what extent the machines are more autonomous. So this simple two diagram, two dimensional diagram, is basically giving a general understanding about how we can map the tools or methods that we can use in that process. At one hand, when we see that actually, the the an an an algorithm is doing the pattern matching and cert like, image processing. And when people are just simply labeling data, we see that there's we give more power to the machine or algorithm itself and then human becoming more an assistant assisting role. It can be other way around where we are the humans are actually taking most of the decisions, and we may be using the data collections in a very simple simple way. And when we move together towards both using machines and humans, certain a large set of solution spaces where we see the machines and humans are taking round, taking turn in in in in in solving the problems. But what could be more interesting, where we can look into as, on the upper and sides and upper right side of the mapping where how we can actually have more cooperative collaboration between machines and the humans. The way when we move more advanced techniques, we we we are very much aware that the complexity increases and also how we are going to identify the complexities of those interactions may become important in understanding whether that we are incurring bias into system, whether we are aware of the implications of the systems, and how what was the data life cycle to that process, where we can audit the the whole process. So, being aware of the complexity of, human machine interaction and identifying the data life cycles, identifying the computational hotspots, and dependencies, could be quite interesting. So as part of that framework, which we are not presenting here, we are also developing a methodology, How we we can starting from a problem description, we can identify work through the complexities when before we move into the design space.

Yeah. If, if you will, would mention the basic components of the the framework we are, thinking and, as a augmentation for a decision and deliberation process, in a nutshell, we are looking into interaction between people, technology, and objects with within the physical world and the digital environment. So it's good to have these boundaries of where it was it was the physical context and what's the digital environments. I think that's that's how we would start to apply a use case or develop a solution space. And within the physical context, it might be interest important to identify how humans are interconnected. What's the social network of it? Are we talking about a neighborhood where the physical proximity is important, or are we talking about a social network which through the family ties, friendship ties, or work work ties? So in the in in the in object side, which we we call silica set that may contain natural and and non natural things. Natural examples could could be trees or bees or or grasslands or forest in an environment, or it's taking any it can simply be a community community library. So being being being able to understand the context, physical context of things and humans, and then moving towards digital environment, we we consider two two representations of humans and and machines or nonhuman entities where where we can name them as digital twins, which can be as simple as data container of attributes, properties, or set of behaviors if it's met matter of humans. Or it can incorporate more advanced interactions, patterns of of those entities. But within that framework, we prefer rather a simple representation, a data representation of humans and things, whether we call them personal digital twins for human entities or asset digital twins for things, then that's that kind of representation in digital environment can feed either experimentation and exploration through simulations, or they may they may feed into software platforms, tools, applications that we do. It can be a smart contract in design. It can be a simple UI design. But what's could be interesting to could be to be aware is that there's a flow of data information to different stages, and that's what we want to achieve in the end, that we create act actionable knowledge. For instance, in that, perspective, a simulation tool can represent how we capture relevant entities within the boundary and then explore that's not possible within the reality of the time, the current time, so we can expand to the future or different scenarios. And then use this as a additional insight in decision making process or learn from these processes and use the output that that kind of exploration either to calibrate or or an existing software or use it as a as a as a kind of blueprint of a certain design. So that these are good to just major components and top down, like, from a general overview of the components we we like to mention.

Yeah. I think here with so we are looking into kind of more coherence of deliberation process as well. We are trying to bring in more than human aspect. So it's I mean, we are we are talking this too many times maybe, mentioning this too many times what what we mean if we have in a community environment. And in in our community environment, we would think that it's important how we kind of expand the green infrastructure. And it might be important to understand when we expand the green infrastructure, which means we are going to share certain resources like by green infrastructure, I mean, maybe a community park or a park that or trees that we want to have on our streets. So that means, well, then we need to be aware how can we look into resources that the non human entities like a tree would need, which could be a water, which could be a space, physical space that we need to all look allocate. And then that requires how we are going to interact with it. And the way we are looking into that kind of being in that that kind of agencies is possible to to actually put sensors and and do the develop upon existing knowledge and research on on the area. And basically taking this research into deliberation representation and deliberation of how a tree as an entity might act as an and has influence on our decisions. How we can actually be able to understand needs and necessities that that entity that may help us to expand our understanding. So here we are so that kind of collective understanding is expanding this democratic deliberation process or experimenting with kind of different kind of the the democrat deliberation process beyond human agency. And when and the other, I think, opportunity we're here here we're kind of mentioning is the when we have awareness to understanding what nature is around us, what they have influence on us, what how we influence badly or negatively, that may if that doesn't go into an action that we commit that may not be sufficient to towards a change. So as part of that aspect, we are that's important to look into what kind of commitments we are having and how we can actually understand the the different types of commitments that that at collective level as well as at individual level, we may be engaged. We can we can engage in. And then without having ways of means of understanding our actions, our decisions, which is through impact model modeling, we will not be able to evaluate change, update our interactions with with with within the society and within the environment. So these three aspects, collective understanding, commitment engines, and impact modeling is kind of essential, aspect when we design a use case for, for community. Yeah. Like, as we have mentioned, so we are talking about challenges of feedback, the ability to measure the impact. I think what we have mentioned earlier, that's important to reconsider, actually emphasize again that when we when we talk about a society or whether it's a group of people, a neighborhood, a city, or a nation in March, we may need to be aware and aligned with the values of the society. So then these are the challenges when you develop a technical systems to be able to provide right feedback, ways to measure the the interactions and influences, but also be aware that there there might be more than one value system. And then at least to have a process where we can actually do the alignment with the the values of the society or the community in the in the process. So from when we from move from problem to the challenges and and and the opportunities is to be able to identify peace and actually support the community.

Yeah. When we talk about, the the climate action or with the support of augmented augmented tools. I guess, we as as we have mentioned, there are opportunities where we can have, expand the agency, which may help, also, our democratic decision processes and, considering more participatory, and discovery, society. And that's may, hopefully, may lead into setting the national or community targets and trying to reach these targets. But when we are in domain of that use of tools, it's it's it's important to be aware of the certain challenges such as are we inclusive enough? Do we have accessibility to all range of individuals, as well as different perspectives? How how we are interacting within the environment? And and that's kind of increase the complexity because we are talking about different diverse context, and then how we are going to, stimulate, diversity to towards outcomes or actions. And an additional challenge as the challenge we are all facing in increasingly important data economy, How are we going to identify suitable appropriate datasets that we are also able to measure the bias that's incurred in datasets, and we are able to audit that bias. But at the same time, we can rely on and we know to what extent we can rely on that these kind of datasets. So we have opportunities, but they're also important to, to address the challenges we we are going to face.

Yeah. Maybe it's before we go to the maybe open discussions, it's important just to mention how what how we can actually address the values and ethics in in in the in the process. Basically, when we look into value alignment, when we look into existing debates, most of those debates are very much technical technological driven. But when we look into, for instance, the the the case of a totally problem, there's certain moral dilemmas. And also, when we talk about values, the the the question is, but whose values are we going to encode in an AI system? So doing through throughout that thinking and framing, we we kind of looked into a number of very good set of research and analysis. One of them is, which is about how we can actually be aware of value sensitive design, which means how we can actually explicitly bring in values, into the design with the support of community. Another source of as a reference is the the efforts within the European Union, which has kind of identified factors and domains of trust for the AI that any application need to address. So and and then and that is also a very good set of frameworks or a set of perspectives that need to be considered. So the the way we are looking into the process is that we need to have a process of value alignment, where people may decide in a in a when we decide, for instance, a a new community park, are shall we put more emphasis on clean air, or would that have impact any of any and if if that would imply that certain side effects either creating wealth differences because of the the pricing around the park that we'll be increasing. How are we going to deal with with this? Shall we also look into ways of capturing that value, distributing to community in general, and how are we going to handle that kind of investment and development process? So a tool like Polis a platform like Polis can help members of the community, even larger society to kind of deliberate and prioritize certain solutions and then use this as a guideline in the developing process. Yeah. For instance, in this if I just meant simply mention the the the domains, we may need to look into human agency and oversight, as well as in in robustness and safety in in the technical implementation, privacy, transparency of the whole process, whether we are and also whether we are aiming and, inclusive processes and accessibility, to what extent social science and environmental well-being is addressed, and how how to what extent the process is accountable. These are the fundamental seven seven domains that, explicitly been needs to be addressed and discussed, and and basically decided how it is it has been incorporated in a in a design process.

Yeah. I guess the the questions are listed in the the chat area. Yeah. Feng, possibly you didn't have the chance to look into. I might start from the last one. So first, the last question is, is there an assumption that the outcomes of these deliberations and discussions will be automatically enforced and implemented, and who's responsible for implementing the collective will. Well, the that's a quite a lot a question that we can discuss in larger quite in detail. But the short version would be it should be part of implementation process if we are going to commit to certain actions. And as part of that process, if the community decides we agree that if 70% of the community agrees into that, let's say, composting collective composting, then that can be enforced or it can be voluntary. I think we can't make an assumptions for the community, but at least saying that we may have we need to have processes in place to be able to deliberate that and then decide whether they should be enforced or whether it should be voluntarily engaged. So I will read this another question from Joshua, I guess. Yeah. So it's like so it sounds a bit that here, like, collective climate action is focused on developing processes for, participation in more private decentralized communities as opposed to participation in existing national political processes. So how would you respond to people who argue that private, individual responses to climate change or even local neighborhoods responses ignore where most of climate's impacts are, from, namely from industrial and agricultural policies? I think that's

people

Yeah. If I may, and I think this is an excellent question, because, we we all know, that the the climate action cannot be solved at individual level or at very low community level. And we we we need a collective global action. I think in in that sense, it's very, very important question. But we also we need to be clear that the the perspective we are looking into, the the kind of augment augmented intelligence framework we are looking into is at community level. So it can function, it can be expanded. I think we can if you can expand this process to national level or using this kind of augmentation to have a global decision making process. But we have more focus on community level with an assumption that individuals are not sufficient, but most of change starts to happen at the community or neighborhood level, and then how we can support that process. So but, we and we we it's important to acknowledge that, we need global action.

I I guess I'll just add, from my experience so I ran one of these experiments in Boston with the the mayor's office there, where we built a lab, you know, did a prototype, collected all the data. And the goal was this sort of, like, to have this, you know, hyperlocal test pipe and sort of provide data to, let's say, like, a broad range of stakeholders. But what we found and probably perhaps what you're finding is that, you know, the people who actually consume these datasets are, like, very specialized. Typically and there's, like, essentially, like, what they're useful for is, you know, like, for, like, you know, planners or people in the mayor's office and not necessarily just, like, average stakeholders, average citizens walking around. So I think it's it's what she was describing here, trying to provide use AI in these datasets and for collective intelligence seems but it's it's quite a challenge. Sorry. Didn't mean to

Since we're so short on time, do you think we have time for maybe one more quick question and then wanna wrap it up? I think you can skip mine. I think Daniel had a question around, what might be the driving force in the transition from the current modes of operation toward this mode that you're describing now? And what are the first steps of this transformation? Or where are these steps happening?

Yeah. I, yeah. Thanks, Daniel, for the question. I think we can the the the the based on the framework we are looking into, it's more tailored to community level action, which means the first steps will be having a community in place, deciding their tools about prioritization, what they can do, what will be the norm of contribution and commitment, whether to what extent it could be enforced, having this deliberation process. And then doing this, what are the values, priorities? How are we going to resolve that? And then making this clear and leading this into either technical or business or investment process that can actually be implemented. But I think the steps that that we're thinking on is not necessarily have to have a huge investment, a huge technology. It can be a very as simple as an analog decision process or an analog another action that collectively we we agree. So the in that case, augmentation can be simple as simple as using a a tool to elaborate on deliberation, or it can be a tool where you collect data or measure the impact. And I think that's that can that kind of spectrum needs to be in place. So we are we are trying to be careful here. We don't need to say that we have to have full fledged tools, expertise to start with. It can it can be a spreadsheet, a table that actually can facilitate the process.

Great. Thank you. We're actually right at time. And I wanted to leave enough room at the end for a tradition that this group has where we well, first, actually, I wanted to is there any other including thoughts anyone had? I also realized at the beginning of this, I skipped the opportunity for to have any new members introduce themselves. And I think Bowen, did you mention that this was your first question?

Yeah. Nice to see you. And, thank you so much for the speakers. This is really interesting and really very thorough, so I'm looking forward to reading more about the project as as it develops. Thank you so much for your time. So could I ask everyone to unmute their mics briefly so we can give a quick round of applause?

Yeah. Thank you for opening up the space for this conversation. Yeah. It's it's lovely group, and it's interesting group. So I guess we'll be more in touch and have other opportunities to discuss. I have seen a very good question about agency, how it's represented. I I would love to discuss that. And, yeah, there are also more interesting questions, but, hopefully, we'll have other platforms or other excuses to go over these details. Yeah. It's a moment.

I'll send you a link to the Slack if you're interested. Or if it's okay with you, I can also share your email with the group. But really Yeah. Your work. So I would really love to stay in touch and see how this how this evolves.