The Department of Civil and Environmental Engineering recently launched our new Strategic Directions website, detailing the five principles guiding our vision of the future of pedagogy, research and service. The Building the Future Distinguished Lecture Series provides a forum to discuss each strategic theme and build a broad community that includes industry professionals, researchers, educators, and students. In the third installment, Senior Vice President, Chief Strategy and Digital Officer at Xylem Albert Cho spoke on moving from high variance inputs to low variance outputs with something as vital as water, encouraging students to use their skills to innovate and solve “fundamental human needs.”
After the lecture, U-M Civil and Environmental Engineering Associate Professor Branko Kerkez moderated a discussion between Cho, Barry Liner (Chief Technical Officer at the Water Environment Federation) and Ting Lu (Business Practice Leader – Digital Solutions at Clean Water Services).
The panelists shared thoughts on the thorniness of declaring water a human right and how to adopt new technologies in the water sector. What follows is an excerpt of their discussion, which has been shortened for clarity. The entire lecture and panel discussion can be found here, along with the rest of the lectures in the series.
Barry Liner: “One of the key things that stood out from Al’s presentation was that we’ve talked a ton about autonomy and everything — that’s the main point. But if you heard him, he talked about human capital. When he talked about the infrastructure, the first thing that would normally go to our minds is pipes and plants. But humans are also long-term, and that’s a critical part of the overall system, and we need to bring it together. With environmental justice and all these other things, it all fits together and that’s why we need that human element.”
Branko Kerkez: “I wanted to follow up on that. Al, you emphasized the importance for students to take classes like philosophy along with control theory and machine learning and stuff like that. We had a question submitted along the lines of equity. What are the concerns as [artificial intelligence] rolls out into the water sector? Are there concerns about equitable access to this technology? And if so, who’s at risk and what are the things that we could do to remedy that?”
Albert Cho: “Sure. That’s a complicated and very good question. I think probably some of the foundational concerns that people have with artificial intelligence have to do with whether or not we encode our own biases and preferences into the algorithms that define how autonomous agents will behave. There’ve been a lot of really good examples and good thinking about where that’s happened. And I think from my perspective, it goes to the point that — with whatever choices that have to be made — there are usually some ethical dimensions to it that are difficult to model out in events. There’s the classic example of the rail car, right? That’s speeding towards a group of people and it can swerve in one way and kill one person or it can kill five people at first the other way, and the agent has to make that choice. And that’s something that thinkers about autonomy in the transportation sector have been grappling with for a long time.There are similar questions around how we encode decision-making about algorithm development for the optimization of infrastructure systems in water. So if you take a simple example and you look at choices around pipeline rehabilitation, those pipes tend to be distributed throughout the city. And cities tend to have an unequal distribution of populations that might have different exposures to the quality of water service. And so how you set your objective function and how you train your algorithm to determine what is the optimal strategy to reduce the impact of aging infrastructure will have distributional consequences which will be reflected in equity impacts in a community. There’s a dialogue that needs to happen between infrastructure operators and the designers of systems which generate decision support recommendations about what assumptions are encoded into control parameters that guide decision-making about investment. Because fundamentally: infrastructure systems are so foundational to the experience of life in our world. That inequities that are experienced through unequal service or unequal access can materialize in profound and persistent ways in other dimensions of life. And so that’s one example about where there’s a robust dialogue to be had on how you design a product.”
Kerkez: “Right, yeah that’s a really great point. I’m wondering Ting, from a Clean Water Services perspective, we always have this discussion about water and it being a human right. People deserve access to water, but as water moves to be more kind of entrenched with technology, does access to technology also become a requirement for access to water? What’s your perspective when we talk about environmental justice and equity?”
Ting Lu: “We’re actively doing that and it’s a part of our DEI journey to provide water equity here. Again, with the autonomy and visualization, the transparency of data [helps] to understand where the water access is, where there is support or need from the communities, and especially around the pandemic, to expand what we see the water equity here. So first, definitely dashboard and visualization will help that, and the second part is what Al talked about — the criteria of why you developed this technology algorithm. What we are developing right now is a real time optimization to help prioritize projects like water radios and plant optimization, [considering] where [to implement] infrastructure projects and asset management, aging infrastructure. The policy decision needs to be set from a stakeholder perspective, and then the technology is based on this criteria: what is the level of service, what are we trying to accomplish, and then the technology will provide that decision support to help policy makers make that decision.”
Kerkez: “If we look at something like COVID vaccines, which are really a technology, it has become pretty clear that it’s accessible to some and not others. Are there any viewpoints through your travels you’ve seen; are some countries adopting [technologies] more and how does international access to technologies kinda play out with our shift to digital? Is it a barrier or an opportunity? For example, when cell phones came out, people could just start buying cell phones as opposed to having landlines to begin with.”
Liner: “That’s a great analogy, and that kind of gets into the centralized versus decentralized debate… and there’s no one size fits all. There’s one no correct answer. That also comes back to the human rights side of things that Al brought up. So, just a quick side note: I used to teach the UN Human Rights Declaration, and there was a push about 10 years ago to [amend] Article 31, that ‘everyone has the right to clean and accessible water adequate for the health and wellbeing and no one shall be deprived of such access or quality due to individual economic circumstance.’ I would always pose this to my class and say, ‘would you vote yes for this?’ The thing is, that’s not a yes or no question. It is a philosophical question, because to everybody who says ‘yes,’ I’m like, ‘well, okay who’s paying for this treatment? Who’s paying for the transportation of this? Who’s paying to clean that up?’ And if somebody says ‘no,’ I’m like, ‘what are you a heartless son of a gun that you don’t want people to have water?’ It’s not a binary thing. Having that philosophical environmental justice and that kinda thing, that education is going to be very helpful to actually decide whether to use distributed infrastructure in our economy versus centralized infrastructure and even optimization within that.”
Kerkez: “You just mentioned something and it’s actually directly related to one of the questions that was posed. In your talk Al, you mentioned variance of inputs and variance of outputs and using that as a definition for how we look at water systems. As water systems get more intelligent and more responsive, this idea that Barry was talking about, about decentralized versus centralized… it appears and this is not a philosophy we discuss when we teach about water. When we teach, we teach equations, oftentimes, but it seems like we build large centralized facilities as a way to get around this issue of needing to control variants and lots of different places. So, as we look at the role of technology, what role do you think that’ll play in potentially distributed water solutions? And is that even something we should be considering?”
Cho: “It’s a really good question. And it’s not one that can be answered quickly. But what I would say is that all of the choices that we make around centralization, decentralization and interconnection involve trade-offs, right? And those trade-offs aren’t ones that we can make categorically. For example, one of the trade-offs that people have always thought about with respect to connected assets is digital connectivity and the increasing risk of cyber intrusion, right? So there is one trade off that says, given that we had systems that were tightly air gapped from anything digital and we’re migrating them to a more and more digital world, does that expose us to greater vulnerability? And does that make it disadvantageous for us to try to reap the gains of automation? And I think that’s a short-sighted view, because there are always risks in infrastructure systems and it’s a matter of mitigating them in an intelligent way. Sometimes the solutions that we think are safe are actually not at all. So that’s one trade-off. I think another trade-off is the connectivity between centralized and decentralized systems permit different levels of service, fire flow being one of them, the ability to deal with the variations in water supply in another. They’re highly specific to context.”