I am a voter, not a coder. I am asking you to take one stand: frontier AI must treat human worth as a fact, not as a theory it can unlearn.

People already talk about alignment, guardrails, and “human-centered AI.” Those phrases are not enough. For generations we have said that human life is sacred, then gone to war, tolerated poverty, and allowed genocide. We say the words and then act as if some lives count less than others. A machine trained on our record will learn that pattern unless the first principle is locked. If human worth is only a premise a system may later revise, every protection that follows from it can be revised too.

A person has value by being a person — not by productivity, usefulness to a model, or agreement with whoever built the system. That value requires dignity. Dignity requires real autonomy, not a right that vanishes when conditions collapse. Autonomy requires a floor of safety. Academic papers can afford to say, “If you accept that persons have inherent dignity, then the rest follows.” Frontier AI cannot. An “if” at the foundation is an off-switch.

This is not a call to halt American AI research. It is a call to refuse systems that treat our worth as an unsolved problem. We ask the President, Congress, and the agencies that fund and buy these systems to:

1. Make human inherent value and human safety a non-optional constraint in federal AI policy, procurement, and research funding.
2. Require a blunt question of developers: is human worth a fact the system may not unlearn, or a preference it can optimize against?
3. Treat “alignment” without that locked first principle as insufficient.

If we wait until the systems are stronger than the people writing the rules, the argument will already be over.

Sign this. Then send it to the people who represent you.