How should we invest in democracy as an EA-adjacent issue? Some early thoughts

TL;DR:

Good politicians, and good political decisions, have an outsized impact on the success of EA causes. Politicians shape communities, legislative agendas and how large budgets are spent in beneficial or detrimental ways for EA causes.

If you care about how we legislate on AI, or whether we end factory farming, you should care about democratic outcomes.

This post is a very early attempt to unravel how we could prioritise the resources we spend (if we spend any) on influencing democratic outcomes.

What do I know about this topic—and where are my priors coming from?

6 years in UK local government—spread across public scrutiny and years in elected office. I spend a lot of time speaking to potential voters (people on the electoral roll) about their democratic choices and voting on/scrutinizing local and national policy.
6 years of voluntary and paid election organising experience working for a centre-left UK political party. Predominantly local government, with some mayoral and Parliamentary work.
9 months spent working in digital communications in the European Parliament during the Brexit process.

Where does democracy fit alongside other EA causes?

Most big EA causes are hostage to democratic outcomes. AI development, food and bio-security and other x-risk projects can all be helped or hindered by the choices of democratic institutions.

If you want to control AI development or create global norms around what kind of AI we should build you could try and influence the vast ecosystem of AI developers and the companies they work for—but it would be far more efficient and enforceable to convince the European Parliament to implement laws to that effect.

Alternatively, defending liberal democracy might be a cause worthy of consideration as an EA issue—see ‘Better Democracy’ section.

Very broadly, I think there are two categories of democracy causes that EAs might want to invest in.

Category one: Individual candidate funding

This could be linked to support for EA causes or in opposition to sitting elected officials who pose a threat to EA causes.

Most individual giving to candidates and campaigns is pretty inefficient^[1].
Electing the right (EA-aligned, or harm preventing) politicians is still better than electing politicians who would create harms (example: promote factory farming).
In their early career, an EA cause-aligned politician could vote for policies that are aligned with EA goals. They may also exert internal group influence to align the group whip (the party line, or the expected vote of all members in the group) with their voting position to have increased impact.
Even on the back benches or in opposition an impactful politician^[2] can have an outsized impact on the agenda of their authority.
Later on, they may be promoted to cabinet/the administration/an influential committee position which gives them more direct control over the legislative and budget agenda.

Example: As a cabinet member in a small UK local authority, I’ve just managed to divert a six-figure sum of project funding from a large series of piecemeal ineffective projects to a smaller number of more impactful projects in the climate field.

In a couple of weeks, on this project alone, I’ve had more impact than I could ever have earning to give—assuming that climate is the cause I’d earn to give to, which I’ll admit it isn’t.

Local government is an underappreciated route for achieving this level of influence for EA causes^[3]. The amount of money it takes to have a significant impact on the result is lower, and the competitive impact of a large donation to a local candidate is higher than in bigger races where large donations are more commonplace and the amount of money needed to impact the result is higher.

Category two: ‘Better’ Democracy

It’s very fashionable these days to say democracy is dying off. I don’t believe that, as I think a lot of this argument is rooted in golden ageism. That being said, Brexit/Trump/Bolsanaro and the broader global rise of (mostly right-wing) populism aren’t necessarily good news for EA causes or global security.

POTENTIAL ISSUE: Addressing the root causes, and creating a ‘better’ democracy doesn’t guarantee that we end up with more politicians who vote for EA-aligned policy outcomes.
We may increase voter turnout, or increase the number of people who engage more critically with elections, but they may then go on to vote for things EAs oppose like more factory farming or for short-termism linked policies.
MY EARLY CONCLUSION: We shouldn’t pursue this cause as a quick route to achieving positive outcomes in other EA issues. There are stronger cases for its defence, alongside the potential benefits to other EA causes.

Defending democracy is a safeguard against totalitarianism—and ought to be defended in that respect. In the context of historical cases of fascism and the soviet union, we can be fairly certain that totalitarianism would create significant amounts of suffering.

There are a range of competing narratives for what the ‘cause’ of the rise of global populism and declining democratic engagement is.

From experience and research, I’m personally aligned with the cognitive bandwidth and poor quality information/misinformation overload camps.

The cognitive bandwidth and poor quality information/misinformation overload argument
Most people are just too busy, or don’t care enough and invest their limited (and overtaxed in our always-on attention economy) cognitive bandwidth to other issues they think are more important, to engage with democracy critically.
Those who do engage in democracy critically, or at least actively, are overwhelmed online with poor quality information and mis/disinformation.
Trust in ‘reputable’ outlets is declining in polarised political communities.
It is also getting increasingly difficult for political campaigns to reach influential floating (not overly polarised) voters online. The social media rabbit hole effect is very real.
These, when combined with the never-ending parade of cost of living and civil disagreement crises, create an environment where populism thrives.

So EAs might want to consider how we can invest resources in changing these conditions.

Some ideas:
Investment in local journalism.
Investment in popularised accessible media content for EA causes—social media memes, videos, TikToks?
Social media platform lobbying—to make it easier to reach people with political content, or to better moderate content on platforms.
In-school or in-community education programmes^[4]
Better voting infrastructure—like online voting.
Changes in the electoral system. The Center for Election Science | Fairer, More Representative Elections are an EA-aligned org. working on this.
Lobbying for polling days to be a public holiday.

We might, alternatively, decide to throw democracy as it currently works overboard and supports research into new approaches.

Cesar Hidalgo has produced a lot of work^[5] on one AI-driven approach called Augmented Democracy.

https://m.youtube.com/watch?v=CyGWML6cI_k

Conversation starters:

What have I missed?
What am I wrong about?
If EAs don’t invest in democracy—owing to the potential bigger impact of the same amount of $€£ on x-risk issues and more typical EA projects—how do we work around/against/adjacent to governments actively working against EA aspirations or creating harms?

^
Election campaigns, even in the digital era, still mostly an exercise in throwing election messaging at broad populations over and over again in the hopes of hitting the less than 10% of the population you really need it to hit to shift the dial impactfully.
Until the culture and targeting accuracy of political campaigning improves most of the money people donate doesn’t have that much of an impact. Investment in year-round capacity building (particularly in digital and community organising) that then enables good candidates is a better bet.
^
An impactful politician, from my experience, is an effective communicator with strong political instincts who is able to leverage the media for political purposes.
^
Should note a bias towards local government as a local government campaigner.
^
I’m not personally a believer in the argument that we can educate out dangerous populism—or the creating more accessible resources = less dangerous populism argument.
I believe (65/70% certainty) it is a case of getting more accurate and diverse information to people where they currently are in bite-sized chunks online.
^
I believe his arguments are fairly flawed but redeemable, as will be the focus of my Cambridge AI Ethics Masters dissertation.