How can we employ technology to facilitate the democratic process? Which platforms are more democratic than others? These and more questions are answered by Kevin Esterling, Professor of Political Science and Public Policy at the University of California in a conversation with the RevDem assistant editor, Catherine Wright.
Catherine Wright: You are grounded in deliberative theory — your book, Politics with the People, is about bringing citizens into more direct contact with one another and their representatives. What do you see as the biggest obstacles to communication between those two groups and what makes you want to bring them into more direct contact?
Kevin Esterling: Deliberation can occur at lots of different sites. My own work, along with Michael Neblo and David Lazer in the book, is focused on a particular venue of deliberation like a town hall format between an elected official and their constituents. That is because we are interested in representative democracy and thinking about how to strengthen it. For example, my city council member lives a block away from here, but my member of congress represents 750,000 people.
The way institutions have currently structured communication between elected officials and constituents is inefficient in that most of the messaging constituents send to elected officials is “top of the head” expressive messaging.
For congress, a lot is organized by interest groups. There are astroturf campaigns to click a button to send a message to your member of congress, and congress members are flooded with mostly unconstructive, “top of the head” messaging about what their constituents are mostly angry about. As a result, members of congress do not really read that correspondence carefully.
Because it is not nuanced?
Exactly, because in a sense members of congress already know that. There are no new insights that come out of that to help the member of congress think about how to be a better representative, or what kind of legislation they should sponsor. As a result, if you look in every member of congress’ office (Democrat or Republican), it is set up in a way where you have one side that is mostly staffed by the lowest paid staffers or unpaid interns responding to constituent correspondence. They respond by quickly sending whatever template they have on file that corresponds to that issue. Then there is a separate side of the office, the legislative side, where staffers work on substantive legislation.
In virtually every single office I have encountered those two staffs are separate from each other and they do not talk to each other. That structure makes a lot of sense because the messaging that comes in is not informative to the legislative side of the office.
If you think about it from the constituent’s perspective, if the member is not going to read your email anyway, why bother crafting a nuanced and informed message? It is a waste of your time.
What I work on is trying to re-design the institutions in which that communication occurs in order to change the equilibrium so that constituents actually have a reason to invest time in researching, thinking about, and reflecting on messaging. It means redesigning the institutions, and changing the equilibrium so constituents feel their voice is heard, and members of congress gain insight from their constituents. That is what we work on in the town hall format – to think about how to redesign institutions to induce that more constructive dynamic.
Can you explain more about the Technology Communication and Democracy Lab and how the work you are doing there is starting to address those inefficient dynamics?
With Michael Neblo and David Lazer we worked with members of Congress to design best practices to induce constructive engagement between constituents and elected officials and how to do it using technology. Technology enables communication at scale in a way that in- person communication cannot achieve as well. We came up with a set of best practices so that, if you were going to do a webinar-like call, that is designed to ensure it is going to be the most constructive exchange.
What are some of those best practices?
The core best practice recommendation we have is for a cross section of constituents to be in the event so that all different voices of the community are represented and heard. That is different from technology — most of our lives we are inside echo chambers where we hear the arguments and reasoning of people we already agree with. Those echo chambers also promote extremism and make it more difficult in our democracy to speak across differences.
Speaking across differences is the core to deliberation and deliberation is core to the actual legitimacy of our democracy. The fact that we cannot speak across differences badly undermines the democracy that we live in.
We also want to make sure that reading material is distributed to participants in advance so everyone feels they are informed enough to contribute to the conversation. We recommend that there is a neutral third party that is the actual host of the event. If elected officials or their staff are running the event there is always going to be this suspicion that they are structuring the event to message their own opinion to make themselves look good. A disinterested third party helps make sure that difficult questions come up during the event and that they are answered.
The last thing we recommended in terms of best practices is that the elected officials themselves are present in the online forum either through video or audio. That way constituents actually know they are talking to their member and not just a staff person. It gives constituents a sense that their voices are heard directly.
After that, I thought about the constraints that technology places on what we are able to do. The set of best practices just mentioned works with any kind of underlying platform as long as there is some kind of streaming presence of the member, like with Facebook Live or Zoom. But after having a chance to step back, I saw the constraints that the platforms themselves impose on the constructiveness of a dialogue.
How do you prevent your better designed platforms from falling into the same traps that not just online communication faces, but also in-person communication? It makes sense that you would have an AI of some kind moderating, but how do you create an AI that will be using certain principles to decide what speech is permissible, without reproducing already existing social inequalities and marginalization?
The two things go hand in hand. We think that the better user experience in itself will induce more constructive engagement. You can think of it really simply – if you are ever on a Zoom webinar, for example, and your member of Congress is hosting a webinar, what you find is that the member is able to message participants about what they want to say, but there are very limited opportunities in a webinar for participants to speak back. In a Zoom webinar, you have a chat feature, and you have a Q&A function, and that is it! So, if you are ever on one of those webinars, or a Facebook Live event which has the same thing, what you see is the elected officials messaging about whatever they want to say and then the chat is filled with this invective of people who are angry because they really feel like they do not have a voice. What we want to do is design the interface so that there are far more robust opportunities for people to engage with the question-and-answer format in a more effective way. And that in and of itself should make them feel less frustrated because they are able to express themselves better.
Then the algorithm that we are designing will look for a list of things that make for a constructive conversation and keep pushing that conversation back to meet those aspirations. A simple thing is having a toxicity filter in place — if someone types in a hostile or abusive message, it does not even show up in the session. Also, we are thinking about ways to make sure underrepresented people in the session have their voice heard.
What work is left to do on your project and what sort of digital infrastructure is necessary in your mind to strengthen these democratic institutions?
We still have to deploy our next version a month from now. It will be the first one that will feel like it is at a point where it will have the user experience for both participants and for moderators. Once we finish testing it, and it works, then we will have groups just start using it, without any algorithmic support. Then once we start collecting data and see how it works in practice, we will be building the algorithm on top of that. So those are the next steps for Prytaneum.
What I would like is for Prytaneum to compete with the commercial technology we currently have. I do not know if that is ever going to happen, since we are stuck with the profit-driven social media that preys on our worst instincts to make money. But maybe we can get better at understanding it and being more reflective as users of social media so that we are less gullible about being influenced by misinformation, or we understand that what we see there is not representative of our communities.
And on top of that, in the work that I did with David and Michael [in Politics with the People] we showed that the constituents themselves really value when their member of Congress does tougher, more honest conversation with them. Constituents love that because they feel like the member is being respectful to them. If we can get them to think differently about using technology to communicate with constituents in this more constructive way, and they start doing that, that could be the thing that leverages this new class of technology — if elected officials can buy into it, then it could be made routine and become a part of our digital landscape in a way that just does not exist right now.
Practically, how do we implement this technology? Are legislative regulations necessary?
There are two sides. There is the constituent side and the elected official side’s use of technology. I am interested in both. I can tell you for Prytaneum, because the core requirement is ensuring that the full cross-section of the community is represented, we want the platform to be as accessible to everybody as possible. So, for example, we designed the platform to be mobile first, and the front-end designers of Prytaneum began their design with thinking about the UX on a mobile phone.
Making the technology mobile first really helps reduce the barriers of the digital divide.
Then the elected officials’ side is something I also work on separately. The public sector is just terrible at embracing new technology. How do we get elected officials to think about this new technology as a solution to problems that they have? I was made the Chair of the Taskforce Subcommittee on Technology and Innovation [by the American Political Science Association] and I was tasked with leading a subcommittee that wrote a report to Congress about how Congress can become better at embracing new technology. We looked at the root causes that prevent the House of Representatives from embracing new technology and made process recommendations. Instead of recommending a list of new software for the House to download and use, we said that before any of that you have to fix the underlying processes of how you purchase and deploy new technology.
The last thing I will say to answer this question is that, with COVID and the pandemic, a lot of members were forced to do remote meetings because of the shutdown and social distancing. Almost all of them quickly went over to the commercial technology like Zoom webinars and Facebook Live. When I saw that I thought it was really promising, because at least they were using technology in a way that they were not using it at all before. But then what I worried about was that, by using technology that is ill-fitted to that use case, they will think of remote engagement as itself unconstructive. They could cement in their own minds that remote engagement is unsatisfactory. But the reason they find it unsatisfactory is that they are using the wrong tools. That is what we want to try to solve.
This transcript has been edited for length and clarity.
In collaboration with Oliver Garner.