Why Fairmind

The Problem

Public debate on important issues is dominated by advocacy rhetoric designed to persuade rather than truly inform. Facts are twisted to fit narratives. Nuance is lost to slogans. Citizens who want a fair-minded understanding are left to disentangle substance from spin.

AI assistants should be able to help, and often they do. But trained on the distortions of public discourse, they can echo those distortions. Some are easy to spot, but the subtle ones—hidden in framing and emphasis—are not. And even when attempting to provide balance, an AI can dilute valid arguments along with the invalid.

Citizens deserve better. Democracy deserves better.

Our Solution

Fairmind creates decision guides that present contested issues fairly, accurately, and substantively. These guides:

• Present the strongest version of each side’s case
• Only use verified facts with citations to reliable sources
• Document editorial choices of what wasn’t included and why
• Structure information for both human reading and AI interaction

How It Works

Our guides are designed to work in two ways:

• With AI: Use ChatGPT to explore a guide conversationally. We’ve created a “Custom GPT” that uses our verified content and follows our instructions for conducting a fair-minded decision process. It walks you through the key arguments and trade-offs, helping you form your own viewpoint, which it can then summarize for you.
• Direct reading: Read the full guide yourself. The content is designed to give you an informed, fair-minded understanding, so you can form your own viewpoint.

Realizing AI’s Promise

AI assistants offer unique advantages in helping citizens engage with complex issues:

• Personalized pacing: A conversational experience adapts to how you want to learn—spending more time on points you find confusing, moving quickly through what you already understand, and going deeper where you want.

• Active engagement: Unlike reading, where it’s easy to skim past arguments you disagree with, a conversational process actively presents each side’s case and asks for your reaction—making it harder to filter information through existing biases.

• Interactive reflection: AI can help you articulate your own thinking, explore hypotheticals, and understand how different values lead to different conclusions—turning passive consumption into active deliberation.

• Summary and synthesis: At the end, you get a personalized summary of what side you took and why.

These benefits do not come autoamtically from AIs’ training on public discourse. Highly curated content is necessary, as is carefully designed process instructions. And all of it needs to be transparent, so people can see exactly what the AI is referencing. This is what Fairmind is creating.

Our Current Prototype

Our current prototype is a decision guide for California Proposition 50 of 2025, a ballot measure about redistricting. Even if Proposition 50 is not relevant to you, experiencing the guide—especially in AI mode—is the best way to understand what we are doing.

The Vision

Fairmind is contributing to a new kind of civic infrastructure: a reliable, nonpartisan knowledge layer that helps citizens engage with democracy’s hard questions thoughtfully and fairly. The knowledge is structured so it can be read directly as a webpage or used by AI assistants to help citizens in a conversational, personalized way.

Ultimately, we envision a decision guide for every issue of consequence in public life. We’re at the beginning of that journey, but every additional decision guide contributes real value, both as a resource for citizens and as proof of what’s possible.

Get Involved

Fairmind is a public service project in its early stages. If you’re interested in supporting this work, getting involved as a volunteer, or simply sharing feedback, we’d love to hear from you.