You’re Not Training AI. But It’s Still Making Choices for You.

The World’s Most Important Decisions Are Quietly Changing Hands

It’s happening silently, but swiftly.

AI is no longer just predicting your next click or completing your sentence. It’s starting to make decisions - big ones, moral ones - on behalf of institutions, businesses, and sometimes, indirectly, you.

From whom a bank approves for a loan to how police resources are allocated… from what shows up in your feed to which patient gets prioritized in a hospital triage room… the influence of AI is growing fast, and it’s getting personal.

Let’s not pretend it’s neutral.

Because it’s not.

These systems are trained on data that reflect the loudest voices, the richest companies, and often, the narrowest worldviews - sometimes from elite institutions, sometimes from online mobs, rarely from the rest of us.

That means we’re teaching AI to mimic a fraction of humanity and hoping it does well by everyone.

It’s like writing school textbooks using only the notes of the first row in class - not necessarily the wisest, just the quickest to speak up.

The Problem With Letting Machines Learn Morals From the Internet

We live in a world where knowledge is becoming decentralized, but power is becoming concentrated. AI reflects this tension perfectly.

Even well-meaning developers can’t fully prevent skewed outcomes when the underlying data represents only a thin slice of society. And for all the advances in training techniques, there’s no technical patch for missing moral input.

Worse, we’ve accepted a broken book of ethics - one that often speaks from a pedestal, not a sidewalk. Every bestselling self-help book has a tone: “Here’s how the most successful people do it.” But what about the lives of everyday people? Their dilemmas, their priorities, their contradictions?

That’s not noise. That’s real life.

And real life is where AI is now being deployed.

Everyone Gets to Vote in Elections. What About AI?

At least in a democracy, you get to vote for your leaders. But with AI, you rarely get to vote for what it learns, how it behaves, or what values it should follow. Yet these systems are becoming global decision engines.

The need of the hour isn’t just safer AI. It’s more inclusive moral input.

AI shouldn’t just learn from data scraped off the internet or from the preferences of a small set of contractors. It should learn from people. From you. From everyone you’ve ever known. Because decisions made at scale demand representation at scale.

How Do We Do That?

1. By Asking the Right Questions

We’re building simple, scenario-based experiences where people can respond to real-world dilemmas - the kind AI is already having to weigh in on.

For instance:

Should an autonomous car prioritize its passengers or pedestrians in a crash scenario?

If a chatbot is asked for politically sensitive information, should it stay neutral, warn the user, or redirect?

Each time someone answers, we learn a little more about what people value - in different contexts, under pressure, and across cultures.

These aren’t abstract philosophy debates. They’re design choices for the algorithms already running our lives.

2. By Making Participation Frictionless, Anonymous, and Safe

You don’t need to have a degree in AI or ethics. You don’t need to write a paper. You just need to show up - and answer with honesty.

We don’t collect personal identifiers unless someone opts in. Responses are anonymized by default, and our systems are being designed so that even the smartest AI won’t be able to trace value inputs back to individuals.

We treat data privacy like a first principle, not a legal checkbox.

Your values are yours. We’re just helping collect and organize them, respectfully.

3. By Honoring the Power of Pluralism

We know that people don’t always agree. That’s not a bug. That’s a feature.

One of our founding principles at Sanctity is that disagreement isn’t a liability - it’s a map. By understanding where values align and where they diverge, we help AI navigate complexity instead of flattening it.

This isn’t about forcing consensus. It’s about mapping perspectives - so AI systems can reflect real-world nuance instead of artificial uniformity.

What We’re Building - And Who We’re Building It For

Not Just for Our Own AI. For Every AI.

Some organizations are doing commendable work in encoding values into machines. A few are even testing ways to involve the public. One example involved asking a group of citizens to vote on core principles that should guide AI behavior.

It’s a meaningful step.

But what we’re building is not a feature for a single product. It’s an open foundation designed to serve all AI - regardless of who is building it.

At Sanctity, our focus is to become a global repository of human values, collected with care, stored with respect, and made usable by any system seeking alignment. We don’t want to train just our own AI to behave better. We want to enable a world where any AI can act in accordance with human values - drawn from real people, across cultures, with transparency and choice built in.

We believe pluralism isn’t a hurdle. It’s a superpower.

We’re Playing the Long Game - But We’re Starting Now 

Sanctity isn’t about instant solutions. It’s about building durable systems that ensure AI behaves in ways we collectively find acceptable - not just technically, but morally.

We’re designing for scale, but also for integrity.

That means:

Making it easy for people to contribute their judgments through interactive experiences

Using strong anonymization so no one’s input can ever be traced back

Exploring how we can compensate contributors meaningfully, especially early adopters who help shape this foundation from day one

We don’t yet have all the answers on how value contributions will translate into economic benefit - but we’re committed to making it fair. In our view, no one should feel their voice was taken for free to enrich a system they never benefit from.

We’re also building toward a world where contributing to the moral training of AI isn’t just a civic act - it’s a job, a gig, or a calling.

For the Thousands Who’ve Already Reached Out

We receive hundreds of emails every week. From engineers, students, parents, researchers, artists. Most want to contribute. Few want to challenge us. Some are simply intrigued.

To every one of you: we hear you.

We’re still laying the groundwork. We’re still refining the engine. But your interest fuels us.

The goal is simple - and profound.

In a world of billions, your values should matter. Your voice should count. And the systems that are learning to make decisions should learn from you, not just from institutions or headlines.

We’re here to make that happen. And this is just the beginning.