Human understandingunderstandingat the decision point.

Modeling human attention, emotion, and behavior can do real good — stronger education, healthcare, and accessibility. Used carelessly, the same understanding can exploit the very vulnerabilities it learns to see.

We build to support people and strengthen their capacity to make free, informed decisions about their own lives — never to manipulate it.

We build and validate our synthetic personas only on human qualitative data. We do not use synthetic data to create them or to judge them — our benchmarks are human benchmarks.

Our test for any output is simple: if it does not reflect something true about real people, it is not good enough.

Coridors are not real people. They should never be presented as real participants, used to fabricate testimonials, or treated as a substitute for lived human experience.

When synthetic output is used, its source should be clear.

Coridor models patterns across whole populations, not copies of specific individuals. Personas should never be used to impersonate real people or imply that a response came from someone who did not give it.

We design against re-identification, so our outputs cannot be traced back to any one person.

Coridor surfaces insight for people to weigh; the decision stays with a person. For consequential choices, there should always be a path that ends with a human.

Synthetic output should never become a way to avoid judgment, review, or accountability.

Coridor can help teams move faster before fieldwork, but important decisions still require human input, evidence, and review.

We do not believe synthetic research should be used to manufacture false certainty.

Synthetic personas can help teams explore perspectives that are hard or expensive to reach early — but they do not speak for real communities. They are a complement to real engagement, never a replacement for it.

We use them with humility, especially around sensitive populations and decisions that materially affect people.

What's legal and what's ethical don't always agree. Plenty of ordinary tactics — manufactured scarcity, engineered urgency, the small hooks that nudge a decision before reflection catches up — are lawful and still corrosive to a person's autonomy.

We don't pretend those tensions away. Naming them honestly, even when we can't resolve them, is part of how we want to be different.

Some uses we will not serve. We reserve the right to refuse work that risks clear harm — even when it breaks no law — and we hold that line against short-term demand.

We keep our own internal standards for where those lines fall, and we expect them to grow as we learn. Where someone believes we have judged wrongly, there will be a way to ask us to reconsider.

We have a defined role in a larger system. Setting the rules for that system belongs to regulators; championing causes belongs to activists. Our part is to build well and be a good steward of what we make.

Where problems are bigger than us, we try to contribute something useful — sharing frameworks and recommendations, including with the people who have the authority to set rules.