On Democratizing

Broader access to information is mostly good

In 2017, I started a public benefit corporation with a mission to:

Democratize access to public opinion research to create a more humane and scientific society¹

That company, Change Research, is now nearly eight years old and I am (again) CEO.

One of the questions I have frequently asked myself is perhaps counterintuitive:

Is democratization good?

Ask most people if democratization is good, and you’ll likely get a mix of reflexive agreement ("duh, obviously") and irritation ("it’s offensive that you would even ask that question!").

But if you asked about specifics, you’d see a different picture.

Is it good to democratize access to nuclear weapons so that everyone can have one? Almost everyone would say no.

Is it good to democratize access to airplane travel? Most Americans would say yes — though there may be a few hardcore anti-growth environmentalists who would prefer to see airplane travel be less democratized.

Is it good to democratize access to automatic weapons so that everyone can have one? Liberals would overwhelmingly say this form of democratization is bad; most (but certainly not all) conservatives would say it’s good.

Is it good to democratize access to abortions, so that any woman who wants one can get one? Here liberals would overwhelmingly say yes, while conservatives would be split (but mostly opposed).

Is it good to democratize access to recreational drugs? Here the consensus has shifted over the course of decades. In the 1980s and 1990s, most people thought marijuana should be illegal. Today that has flipped, and democratizing access to cannabis is increasingly viewed as positive. Meanwhile, democratizing access to other drugs is still viewed negatively.

Is it good to democratize access to expensive but effective medical treatments? Almost everyone would say yes; bringing down the cost of effective treatments to make them more affordable would confer great benefit on society.

Is it good to democratize access to tasty but unhealthy foods? I am going to decline to speculate on this one.

What about access to information?

Again, in the abstract, most people would say that democratizing access to information is good. However, in practice, it's more complicated.

Change Research makes polling significantly more affordable and faster. Organizations that previously couldn't afford polling at all can now access these insights, while others can poll more frequently. A political campaign that raises $125K isn't going to spend $40K on an expensive traditional poll; Change Research innovated to be able to deliver high-quality, accurate insights to these small campaigns for around $10-15K.

This democratization has real impact. Katie Muth, running for state legislature in Pennsylvania, said after she was elected that

Some political insiders didn't think I could win. Change Research's fast and accurate polling showed me ahead at 80 to 90% less than a traditional poll. When I won on Tuesday, I came in precisely where Change Research predicted.

But democratizing access to polling also has potential downsides. There's plenty of misleading polling designed to advance particular narratives, and democratizing the ability to ask skewed questions and then spread the results widely is, at best, debatably beneficial. All polling companies, including Change Research, face client and business pressure to frame questions in ways that might be slanted and deliver results that are actively misleading.

This connects to broader debates about information access and free speech. With misinformation and disinformation, the picture gets even murkier. “Democratization” means a broader array of theories can become known or popularized. Sometimes these theories — think Pizzagate or Bill Gates implanting chips via vaccines — are completely absurd. In other cases — think the Covid lab leak theory, or concerns about lengthy Covid shutdowns — the story is more nuanced. Democratized access to information leads to more knowledge about the crazy conspiracy theories that have no basis whatsoever, but also to plausible theories which are too quickly shut down by the powers that be.

Inside companies, reporting analytics teams face similar dilemmas. Should they act as gatekeepers limiting access to unofficial reports? Or should they "democratize access" and risk people generating misleading analyses?

What about AI?

Artificial intelligence presents another domain where democratization brings both opportunities and challenges.

Democratizing AI will likely wind up being mostly positive — making all of us smarter, more organized, and better able to flesh out our ideas (as I did with this post!). AI can help us explore complex topics in ways that resonate personally ("explain quantum computing using basketball strategy as a metaphor") at minimal cost.

However, there are risks that AI could fundamentally change certain societal dynamics — like how we interview and assess job candidates — in ways that could negatively impact our society. We should acknowledge these potential downsides while generally supporting broader access.

Operating Principles Around Democratization

Broadly, I think it’s best to proceed as if democratization is mostly good, but with an eye towards caution:

• Democratizing access to something generally means bringing down its cost and/or increasing its availability. Democratizing access to something is often underrated in importance: many innovations make something possible but only accessible to people with specific technical skills or large amounts of money.

• I start with the presumption that democratizing access to something will be positive; that should be the default.

• If (and only if) there is specific harm that can come from something — e.g., a clear link between increased access to _____ and increased violence — policy makers (governments or companies) might take steps to make access more difficult.

This is both a philosophical and pragmatic position: it's better to default to openness and allowing things to move forward, while keeping a close eye on potential downsides.

The Political Dimension

It's worth noting that while Change Research primarily works with Democrats rather than Republicans, our underlying belief is that leaders better understanding what people want and need is a net positive. This connects directly to our mission of creating "a more humane and scientific society": we believe evidence-based decision-making leads to more humane outcomes.

This is especially true in non-zero sum games — an election between two candidates is zero-sum, so we can really only work with one side, but a Republican mayor better serving her constituents through scientific understanding of their needs can be a net gain for everyone.

In the end, democratization isn't inherently good or bad — it depends on what's being democratized, the context, and how the newly accessible tool or information is used.

I’m going to spend my time building products and services for which I believe increased access (democratization) is a net positive. From there, it’s possible to move forward with the presumption that more access is better than less, then monitor for potential harms. Democratization isn’t always good, but it can be.

1. We now speak about facilitating a humane, scientific, and just society.