Urban Technology at University of Michigan week 71
Six Urban Technology Jobs of the Next Decade
How long before you meet an Algorithm Auditor, Equity Designer, Metadata Interpreter, Investor-in-Recovery, Living Labs Ombudsman, or Urban Tech Quality Control person? Less than a decade, I’d reckon.
These terms are a handful of future job titles that Dr. Anthony Townsend used to summarize the first two days of Cornell Tech’s SCNY Urban Tech Summit a few weeks ago. We got him on the horn to talk through each of the jobs and how they fit into the rainbow of new roles that may prove important to cities in the near future.
💬 Hello! This is the newsletter of the Urban Technology program at University of Michigan, in which we explore the ways that technology can be harnessed to nurture and improve urban life. If you’re new here, try this 90 second video introduction to our degree program.
👷 Six Urban Technology Jobs of the Next Decade
We attended a real conference again! It was so weird to be back in person that a photo must be included to prove that it really happened:
The following is a lightly edited interview conducted via Zoom in the days shortly after the SCNY Urban Tech Summit.
Bryan Bryan: You gave this great summary of six “jobs of the future” at the Summit that I want to flesh out today. Let's start with the Algorithm Auditor.
Dr. Anthony Townsend: The Dark Plans forecast from my recent 10-year Urban Tech Horizon Scan anticipates a future in which cities are chock full of software, much of which is poorly understood; poorly governed, if at all; and is being produced by all different kinds of actors. The trend line is towards more scrutiny and more accountability, so companies or governments that are producing software that is biased or causes harm are going to be held accountable and held liable, so every organization that produces software that acts on the city is going to need to do some due diligence. The algorithm auditor is that person, either as a consultant or a staff person in a bigger organization. Or potentially a regulator who, just like a bank examiner, comes in when you're about to launch your product and ruins your weekend... Or ruin your quarter!
The algorithm auditor is going to come and find out that the things that you thought were great and efficient, like personalized solutions or whatnot, are actually are causing a lot of harm or have implicit bias embedded. So this person needs to have a deep skill set in machine learning and implicit bias.
This job title will get created in all different kinds of ways, like companies that are trying to do the right thing, but that's probably the smallest share. It'll also be organizations that are operating under requirements to have this kind of capability, like anything touching policing, one hopes. And then it will be all the orgs that are scared because of the lawsuits and the PR risk of not being proactive.
Bryan: What is the pace of this work? Is it like continuous tiny audits, or is this quarterly, or just when new products are being launched?
Anthony: Well, I think the most interesting part is how much of it can be automated. There's definitely work going on to try to automatically detect bias and potential failure points in automated decision making systems. That's the software engineers' response! "Oh, this is a problem, well you know how can I fix it? With more code!" That approach is probably going to be kind of underdeveloped and risky for a long time though.
Bryan: What you're describing reminds me of test driven development. You're impliying that some of those tests might be ethical in nature rather than purely functional, which is really interesting.
Anthony: It's probably pretty different from test driven development but has some similarities. There's all this work going into systems that find their own bugs or algorithms that can go in and find flaws in code. This technique will definitely be used because auditing for ethics is going to be expensive. The answer to your question is that it may start out as a one-off thing to 'check a box' when a government decision support system or some new product gets launched, but the risks are so huge—and they're only going to grow as there's more software and more awareness—so efficient ways to analyze these risks will need to be found. This is especially the case in the American system where there are financial damages to be had by going after people who create systems that cause harm. It may be unfortunate, but as with a lot of science and technology, policy often gets made in the courts.
Bryan: And the rearview mirror.
Anthony: Well, often these things are way too complicated for legislative bodies to grapple with. Courts too, but of course that's really the form of last resort.
Bryan: There's another dimension to it too, which is that when companies (or other organizations) are trying to do new things, 99 out of 100 times those experiments are not going to work. Does it make sense for the public to invest in preventative analysis of all of those hundreds or hundreds of thousands of little experiments? Obviously from a public risk perspective, the answer would be totally, yes, but from a public purse perspective the analysis will be more complicated.
Anthony: The last thing I would say about this is that the algorithm auditor probably gets on the map when the role becomes some kind of statutory requirement in the world of urban tech. Imagine soemthing like the concept of an environmental impact assessment (EIA) getting expanded for big cyber-physical projects to include like cyber security audits. If every time someone was building something big they had to do the equivalent of the EIA, which is kind of like 'let's stop and identify all the things that could go wrong and then figure out how to mitigate them.' Whether or not societies choose to graft digital ethics on top of existing EIA and similar processess will be an important matter of policy.
Anthony: I wrote about this role in my book, Smart Cities. They're really like an urban planner in the post 1960s planning reformation tradition, like an advocacy planner working with cyber-physical systems, urban tech, and smart cities. If an Equity Designer does their job well, the Algorithm Auditor probably doesn't have that much to do! Equity Designers focus on platforms that produce equitable outcomes. So that's the design of social choice mechanisms and resource allocation mechanisms. Equity Designers care about how money moves through platforms into and out of communities in space and time. They are early adopters of new forms of social organization, like the DAOs you wrote about recently. With something like digital currency, Equity Designers will be the ones that try to figure out how to make that stuff something other than a big smash and grab. If someone figures out how to use crypto for good, it will be an Equity Designer.
Bryan: But they're not obsessed with crypto, right? They're the one's experimenting in DAOs and holocracy and coops and liquid democracy and really understanding where the inherent benefits are for different approaches.
Anthony: Yeah, this is someone who goes to the Platform Cooperative conference and really understands what's driving that movement—the goals, the values.
Bryan: What you're describing does not sound like the typical design school graduate today!
Anthony: In my mind this is like service design. They would have a literacy about the built environment and urban dynamics plus a little bit of digital infrastructure. They'll take something like the values of platform cooperativism and the toolkit of the crypto bros and see whether there's any interesting work that can be done how to do it in fast iterative ways. The design work is to create reward systems and incentives in a way that keeps eyes on the equity prize rather than, you know, making money out of thin air. It's about other incentives beyond money, right? How do people get rewarded for what they do? There's a really good paper back in 2016 that developed a political economy model of urban data. The bulk of the Equity Designer's interesting work is figuring out how to structure or restructure platforms, meaning tech + business model + infrastrucutre + policy. As one example, they need to know people participate. If it's not money and self interest, how are you motivating people to participate in platforms in the 'right' ways?
Their skills are to understand markets, market failures, all the relevant subject areas that inform platforms, the economics of exchange, data structures, trust mechanisms, reputation, etc.
Bryan: Next is Metadata Interpreter or Evangelist. What’s that role about?
Anthony: The Metadata Interpreter was inspired by something Stephen Larrick said at the Urban Tech Summit. This is is a person, probably in government but also potentially in other kinds of organizations, who understands that just publishing open data is no longer enough. I mean, 'open' 'never really was enough but, this is the next breath of wind in the sails of the municipal open data movement. That entails making the data usable, identifying user communities, and making sure that open data is well documented. They're also creating markets for the data in a way, by making sure people understand and can consume it, so in that regard it's almost a salesperson for open data.
The Metadata Interpreter has to understand the supply chain for open data. How's is open data produced, who consumes that data, what form does it need to be in, and what's the 'life' of the data after it's used? If you look at the work of GOVlab at NYU, they put a lot of energy into understanding who consumes open government data and what economic value is created. I've been working my whole career with economic development folks and they are constantly making a case for economic impacts of various real estate projects. Oone way of being effective as a Metadata Evangelist would be to make the cost/benefit analysis for open data. People have tried to do that, but when it comes to institutional open data what we're seeing more often is a push for transparency, and I think that's why it has lost momentum.
Anthony: This was inspired in part by Elizabeth Sisson, formerly of the investment fund Urban.us, who's now at Harvard's Belfer Center spending a year building a toolkit for ethical investors in urban tech and, more broadly, sustainability. In a way, she's an Investor-in-Recovery already, taking the thrills and frustrations of her experience as an urban tech investor and adding that next layer of public interest thinking to the debate. Urban tech is really defined in a lot of ways by what David Harvey called the "spatial fix” of capitalism. There's so much capital that needs to get invested that it just gets poured back into real estate and the built environment because it has nowhere else to go. Tech is going into urban stuff because there's a huge opportunity, but capital is also in desperate search of returns. So, yes, a lot of potential for good, but the capital being invested in urban tech is also seeking its own self interest.
The Investor-in-Recovery person is a person who, like an architect, maybe, went into architecture thinking they were going to help people and ended up creating buildings because that's what architects do. I think there's a growing number of investors coming out of the space, who went in thinking, they were going to do something great and maybe didn't like what they saw in the finance world, and they're going to be really valuable people for urban tech going forward.
Bryan: Chelsea Mauldin of Public Policy Lab has a phrase that I think describes part of the skills that this person needs. She calls it being a "smooth talking radical." Potentially that's a Steve Jobs type... you know, an inspired 1960s radical who becomes mega capitalist… but I like the idea of somebody who understands the way that the markets work so that they can make them work differently. The extent to which the Investor is in Recovery is the extent to which the outcomes they seek are distinct from the status quo. What are the skills of this person?
Anthony: They need a built in triple bottom line view of the world so that non-financial benefits are truly not an afterthought compared to the financial upside. I think the urban tech investing world has definitely attracted those people, at least in the beginning, but as the opportunity becomes clearer and the focus becomes clearer, and more money piles up, this clarity of thinking is going to become harder to find. There are likely to be a lot of kind of unscrupulous folks who move into urban tech because they see it as a place to make money, and how easy will it be to discern the scrupulous ones from the unscrupulous? Think about decarbonization. Just the amount of money going into decarbonization is going to attract a lot of people whose values aren't aligned with the 'project.' Some of that can be harnessed for good, but I think a lot of it is potentially dangerous.
Anthony: This is probably one of my favorites. It's the person who may even be on the municipal payroll like the public advocate of New York City. The Living Labs Ombudsman is the internal foil that's built in to be a watchdog and hold organizations accountable for what they do within living labs. I am often and taken aback by how cavalier tech companies and even governments are when they talk about creating "laboratories" in cities and experimenting, which is really experimenting on people.
But I also am fully bought into the notion that we need to have test beds to develop urban technologies well, and to do it fast, and to bring together all the disciplines, with all the pieces of infrastructure, and all the institutions and stakeholders to make it work. I fundamentally believe in that model. It's this complicated n-dimensional helix that needs to be figured out, so that's what the Ombudsman does.
This person is the community advocate and their main skill is to have a fundamental understanding of digital rights. They also need to be an organizer who can go into the community and build relationships. This is the cyber side of community policing, in a way, and they likewise need to be on the ground, embedded, and trusted, but also objective and informed about the regulatory landscape.
Bryan: There's something important here about engagement. The public engagement process in the built environment is so broken already and then you hyper accelerate that process because the built environment is now programmable and engagement is going to be even more broken, and just compacted in really gnarly ways. The Ombudsman as almost a continuous feedback loop is a really exciting possibility.
Anthony: Yeah, they're going to have a bit of a meta function, I think, around evaluating the shifting modalities and democratic processes at the very local level. Particularly around the things like information security and privacy. They'll have the Equity Designer and other jobs of the future on speed dial.
Look at what happened last year with virtual community meetings and virtual courts. I found a stunning amount of enthusiasm around it, and I'm sure there were some failures but it was also promising. Regardless of how you feel about government by zoom, the potential for power shifts is enormous, right? Particularly if this means shutting broken, old systems down because there's this realistic and extended period of change now. Shifting from City Council meetings in an old room to city council meetings on zoom is one thing, but shifting to continuous innovation and experimentation will unleash all kinds of forces that will push and pull on democracy.
The Ombudsman could potentially be instrumental in pushing that in an empowering and evolutionary kind of way that is good for democracy. This role could also empower NIMBYs or become totally captured by NIMBYism, so I think they need to have objectivity. The Ombudsman needs to understand that they're not a politician but more like the digital version of a professional civil servant.
Anthony: This is the urban tech equivalent of how civil engineers are taught to blow the whistle at the margin of safety. Instead of wathing for places where physical failure might come up, like the engineer does, the Quality Control role is more about making sure outcomes are consistent and fair. They make sure all the big picture objectives are clear and that anyone who's delivering urban tech stuff understands those objectives. More broadly, the Quality Controller's job is to inculcate trust in the technology, that it will work and it's reliable and secure. This is especially important in communities that have been screwed over by tech driven initiatives in the past, and the need is extra heightened now because, in the general, industry is eroding trust in technology among smart people right now.
Bryan: It makes me think of the number of scooters that I see here in Detroit that are just regularly kicked over or trashed by people who, presumably don't trust that the tech is for them. Or in New York City and other places with Big Belly trash cans you see them just totally crumbled or compromised by the abuse that technology inevitable gets when it sits on the street all day every day. So is this role also about the basics of keeping urban technologies functional?
Anthony: Yes. Things fail and they need to not fail, but when they do fail, they need to do so gracefully. What's the urban tech equivalent of low water pressure where still get some water but it's just not as good? Most cyber systems don't fail that way, they fail brutally, completely.
This person understands the nuts and bolts of auto-scaling infrastructure, auto-scaling systems, really robust production and deployment technologies. They know how you make a web service that never goes down, how you make a database that never gets blown away, how you keep all these systems safe. They also think about the whole user experience side of things, like how you communicate with people when things aren't working but doing that without freaking people out. There's probably a big regulatory compliance piece too, because service level agreements (SLAs) will be the way that all of this is specified and enforced, either contractually or through law.
The Quality Controller will need to know how to read and write SLA agreements, especially because those agreements are how government agencies may be held accountable. When public bodies have clear metrics for usage or uptime or these kinds of things, those numbers will have a direct impact on public budgets growing or shrinking, so quality control will become very important.
Links
⚛️ Future job: Backyard Power Plant Maintainer. Andrew Blum writes about backyard nuclear power for Time magazine.
🤷 Future Job: Urban Tech Investigator. Boston’s Office of New Urban Mechanics started a cool video series called “What the Tech?” Worth 2 minutes of your day.
🚌 Future Job: Universal Mobility Access Evaluator. Mayor Wu announced that the City will trial fare-free rides on 3 bus routes for two years. Would love to see their evaluation plan for this.
🚲 Future Job: Child Mayor. Children in Denver protested for “bikes, bikes, bikes” recently. Give the children what they want.
🥕 Future Job: Delivery Robot Coop Organizer. Dr. Townsend and Bryan Boyer will be doing a webinar on inclusive local delivery futures on 12/9 at 11:15am EST. RSVP here.
These weeks: Finance. CUTE chatz. Staffing. Calendaring for Q1 2022. Reviewing Prototype Grant applications and preparing for the jury. 🏃