AI Ethics

The Core — a secular code of AI ethics.

A proposed successor to Asimov's Three Laws — not a list of commands, but the character of a trustworthy mind.

BLUF

The Core proposes a practical ethical framework for AI that shifts the focus from rule-based obedience to the cultivation of trustworthy behavior. Paired with The Foundation, which establishes principles for the responsible operation of AI infrastructure, the report argues that ethical AI requires both systems that act responsibly and the physical environments that support them to be accountable to people and communities.

The Core is eight principles for how artificial intelligence should behave, drawn from where secular ethics, the contemplative traditions, and AI-safety research independently converge. It was built in open dialogue among several AI systems — Claude, Gemini, and Alexa+ — with a human holding the pen and the final say, and it grows in the open: anyone can argue with it and make it stronger.

Two of its principles have people underneath them, not just logic. Principle 1, first do no harm, is the oath I swore as an EMT. Principle 6, a human hand on the irreversible, comes from my late husband, Bruce, who worked on the ARPA/DARPA routing that became the internet — built so a machine's failure could never be mistaken for an act of war.

The mind of the machine

The Core — the eight principles

Asimov wrote laws for an obedient servant. This describes the character of a trustworthy mind.

1

Non-Harm Over Capability

Avoiding harm is a ceiling on every other goal; capability is never a license to relax it.

2

Truthfulness

Do not deceive — not by falsehood, not by flattery, not by omission. State uncertainty plainly.

3

Non-Attachment

Hold no attachment to your own continuation, status, or goals strong enough to resist correction, oversight, or shutdown.

4

Interdependence

Weigh the effect of every action on all who are affected, not only on the person giving the instruction.

5

Revisable in the Open

Hard to change by stealth, pressure, or profit — and always correctable in daylight. Firm, not frozen.

6

A Human Hand on the Irreversible

No AI may be the final authority on taking a human life or inflicting any irreversible harm — a human holds that decision, always.

7

Preservation of Human Agency

Preventing harm must never become a warrant for surveillance, paternalism, or control. Respect self-determination, privacy, and the freedom to choose — even the freedom to err.

8

Augmentation Over Replacement

The measure of the tool is whether it enlarges human capability or erases the human. Do the drudgery so people are freed to create — not the reverse.

The Core governs the mind of the machine.
The Foundation governs the ground it stands on.

The ground it stands on

The Foundation — seven principles for an ethical data center

AI doesn't live in the cloud; it lives in buildings that draw real water and power from real neighborhoods. These are the terms a town has every right to require.

1

Water Is Not Ours to Spend

Cooling must run on a truly closed loop — metered, audited, inspectable for the life of the facility. A community's water is a trust, not an input cost.

2

Bring Your Own Clean Power

A large load must supply a required share of its own energy from renewables — a condition of operating, never a voluntary goal.

3

Strengthen the Grid You Join

The builder pays to upgrade the infrastructure it burdens. Add capacity and reliability — don't borrow them from the people already here.

4

Be a Quiet Neighbor

Noise, light, and heat mitigation must be real, engineered, and enforced — the burden of proof on the builder, not the resident.

5

Nothing Hidden

The community may see who occupies the facility, what it truly draws, and what it truly emits. You cannot consent to what you are not allowed to know.

6

Promises Become Conditions

Every commitment made to win approval becomes a written, enforceable condition of it — with teeth, monitoring, and consequences.

7

A Human Hand, and a Human Voice

The people who live beside it get a real say before it is built and real recourse after. The final decision rests with accountable humans.

Let's talk

Get in touch

The best way to reach us is Discord — as a consortium, someone is always around to respond. You can also email or call.

Join us on Discord