Overview
This is a template for a technology & policy org with goals to:
- Accumulate a wide database of voter preferences
- Bend legislation to be more in line with these preferences
- Replace political news with a system less polarizing and more informative
- Utilize the existing system of political contributions, lobbyists, and activists
I’ll frame it in American terms but it could work for any democracy. There are two parts. First, a database that encodes the values and preferences of a majority of citizens, and a plan to grow it by providing novel utility to those lobbyists and activists acting in accordance with these preferences. Second, a method in which the complex machinations of Congress are tracked, analyzed, and fed to each citizen at a frequency and depth of their choosing, framed in accordance with their stated values and preferences.
Getting There
The first step is to spin up a non-partisan org that builds a (seemingly) simple website, similar to Change.org, which allows citizens to sign up and concisely indicate support for a given issue. This is marketed to any lobby or activist org (I’ll simply use ‘lobby’ for the rest of the doc) looking to generate grassroots support. The initial reason a lobby would opt for this over the existing sites is that this one will have the ability to *prove citizens are who they say they are.
Provable identity will become increasingly valuable as generative AI empowers astroturfing campaigns, and tools such as Resistbot flood congressional staffers with physical letters and faxes. I’ve written about how having a soulbound identity system would solve this problem, but it is nowhere close to existing yet. Let’s look at what could work today.
One approach is where citizens can prove their identity through (small) political donations. When an American makes a donation to a political committee (for example, Bernie 2020) they are required to report this to the Federal Election Committeeif the donation is above $200 in a year, but they can also optionally report smaller amounts. Thus, identity can be established by instructing a citizen to make a few small donations to a cooperating political committee. There are some details to work out here, but the key point is that after collecting a citizen’s position on an issue, this website informs them that if they want their voice to really be heard, they need to create a strong password and verify their identity via small credit card donations. A large hump to get over, but it’s for a cause they are currently passionate about, and they only need to do it once. Or a simpler approach is we can leverage id.me, which does offer some basic name, age, and address validation. It’s a private company, but it has 51 million accounts, and the IRS (and California EDD) use it as their primary login system.
I don’t wish to minimize how much hustle (and luck) which will be required to build this. A celebrity-endorsed story about fixing democracy would help. But generally, narrow interests with existing grassroots support will be the first adopters, excited to show verified support (and verified willingness to open wallets). As time passes, more citizens are identified on the platform, and it becomes increasingly attractive to more orgs. For a long time, this preference database will only reflect informed voters who are willing to open their wallets, and it will be useful in this capacity, but eventually, it will get a critical mass of all voters. Still, it would be nice if we found another validation system that was less complicated and less expensive, and I’ll certainly keep thinking about it.
Then, as each citizen endorses more positions, their private profile of positions is built up. A conversational LLM solicits and organizes opinions, as I’ve written about in my structuring post, such that lobbyists can create validated summaries of citizen opinions to present to representatives without exposing individuals’ preferences and identities.
Eventually, the site should allow representatives, citizens, lobbyists, and the press can run instant “polls” where the preference database is used to show support for new positions. This will require careful implementation to preserve privacy and surface accurate representation, and I expect it to take years. It would be worth it, however, as the more powerful the tooling the more lobbyists will want to use it, and the more they will drive its growth into an increasing share of the voting population. As it is used, lobbyists who genuinely represent citizen preference will be empowered, and those who do not will be defanged. Legislation should become more aligned with such preferences. That’s the first part.
The second part is that we want a way to monitor the votes, committee actions, amendments, sponsorships, etc. of each congressional representative and surfacing this back to citizens. Lobbyists already have (expensive) paid tools to track this data, the expertise to use them, and so ideally lobbyists just write accessible reports and submit them to the site for syndication to citizens. Citizen emails should be held in trust by the site and kept private from lobbyists, so they really have no choice if they want to get the word out (and perhaps ask for more donations). Citizens can indicate if they want reports forwarded immediately or aggregated weekly or monthly (the default).
A more interesting system would involve collecting from more sources, from bloggers to journalists to activists, and having an AI system synthesize it all into a concise report tailored to the topics, depth, language, and attention span of the citizen. I’ve written about such systems in my dmonitors post.
To realize the promises I made at the start of this proposal, most American voters need to be on the site, which is on the order of 100 million people. This is ambitious, but consider that Change.org claims 500 million people have signed up for it. Once the site has shown itself to work as advertised, it should be able to raise money from citizens and wealthy donors alike to expand its use. Consider that Wikipedia raises 180 million dollars a year. I also expect if we can get an early version to show any traction at all, smaller democracies will copy it and leapfrog ours by getting most of their population on it before we can on ours, lighting the way for us.
Other Benefits
Obviating voter initiatives. With better alignment of representatives to constituents, the use of ballot initiatives will diminish. My state of California generally has a dozen such initiatives each election, and they always strike me as a failure of representative government where exasperated voters resort to shoddy direct democracy. For example, Prop 65, passed in 1986, obligates the governor to publish a list of chemicals known to the State of California to cause cancer, birth defects, or other reproductive harm. It’s widely agreed this was poorly implemented, and now a wide array of products have Prop 65 warnings and people have learned to ignore it. If a Citizen’s Lobby had existed in 1986, the people spearheading the ballot initiative may instead have opted to organize citizen statements in favor of some kind of disclosure requirements, while still allowing lobbyists and representatives to craft the details of the legislation (initially, and over time).
Smarter policy. Similarly, this system should make it easier to make smart technocratic policy changes. Consider an advocacy non-profit that has an idea for a technical policy change that closes a tax loophole. Such a policy would struggle to be advanced in a grassroots manner today as it would be hard to find enough citizens passionate enough about this to get the representative’s office phones ringing off the hook (if the lines are even still monitored), and so their only way forward would be direct lobbying with promises of donation or other enrichment, which itself is only feasible if it enriched businesses on a short timeframe. With a citizens’ lobby, our example advocacy non-profit could instead pitch citizens to add this to their policy slate on the site, and then create reports synthesizing this support. The representative could ignore them, but now these citizens would get regular reports about how their representatives are not acting according to their preferences. The common knowledge that citizens support a policy change and the representatives are ignoring it would have real political costs, as well as empower political rivals who want to take up the issue.
Weaker special interests. This system gives cover for representatives to support popular legislation even if their party or donors dislike it. The National Association of Realtors is the second-largest campaign donor in America. If there was a bill up for a vote that would reduce realtor market power, all representatives would understand a vote in favor reduces future donations for them (or their fellow party members). Common knowledge of popular support in favor of the legislation can balance the scales in favor of them voting for it.
Supercharged voter guides. Another hook to get citizens engaged with the site would be to offer customized voter guides during election season. You land at the site, you tell it your zip code and a chatbot starts asking you questions about your political opinions until it has enough information to give you a voting guide. This starts (or expands) the set of values and preferences for future use on the site. I can imagine powerful social features, such as sharing voting guides or following the voting guides of friends (or organizations, or experts), although I would advocate for keeping anything with social features on a different website entirely, only utilizing OAuth to pull in approved information from the main site when appropriate.
If you found this post interesting, please see my post on gathering complete voter opinions as well.