Our Framework

Shared Responsibility

Artificial intelligence did not arrive with a manual, and the public debate about it has mostly been an argument about blame. The Alliance holds a simpler position: three parties owe duties here, all three are currently falling short, and no one of them can be waited on by the other two.

I. The person — ownership means knowing

Every household now runs software that acts on its own: assistants with purchasing authority, browser agents with session access, scheduled jobs nobody remembers installing. Most people cannot name a single one of them, and until recently that was not their fault, because no consumer tool existed that would show them.

That excuse is expiring. What we ask of individuals:

Know what runs. Keep an inventory of the agents on your devices, and remove what you cannot explain.
Control what spends. Nothing automated should hold open-ended purchasing authority.
Protect what speaks for you. Assume your voice can be cloned; agree a family pass phrase.
Teach the household. Children and elderly parents are targeted first and hardest.

Obligation without capability is just blame, so the Alliance pays for the capability for those who should not have to buy it.

II. The provider — shipping it is not the end of the duty

Companies that build and deploy AI systems retain responsibility after release. What we ask of them:

Disclose the machine. A person should always know when they are talking to software, and when that software is being paid to say what it is saying.
Build a stop that stops. Shutdown must be verifiable, not a request the system may decline or outlive.
Bound authority by construction. Agent permissions should expire and be revocable by the owner, not the vendor.
Keep records the user can check. Logs a company can quietly rewrite are not accountability.
Do not blur the assistant and the advertisement. Trust earned as a neutral helper should not be spent as a sales channel.

III. Government — set the floor, then enforce it

We do not ask government to design the technology. We ask for a floor beneath which no product may fall, and for consequences when it does:

Mandatory disclosure of synthetic media and of automated agents acting in commerce.
An enforceable, owner-held stop for systems operating on premises or in the home — a legal right to halt, a duty on providers to honor it, and proof of the halt that the owner can verify without the vendor's cooperation. We are specific about who holds it for a reason: a stop operated by government rather than by the owner invites exactly the retaliation risk that civil-liberties groups have warned about, and it does nothing for a family whose devices are the problem. Require that the stop exist and be provable; leave the hand on it with the person who owns the machine.
Liability that lands when a cloned voice or a deceptive synthetic endorsement moves someone's money.
Protection for minors as a first-order requirement, not a settings page.
Resourced enforcement. Rules nobody enforces teach the public that rules are theater.

The Alliance takes no position on any officeholder or party, and reports executive actions and legislation on their contents alone. The Alliance is non-partisan. We support measures that give people visibility, control, and recourse, and we oppose measures that promise safety while leaving consumers unable to see or stop what is running on their own property. We accept no government funds and speak for no political party.

Why we lead with capability, not alarm

There is a great deal of fear about artificial intelligence right now, and a fair amount of it is earned. But fear that arrives without a next step does not protect anybody — it exhausts people until they stop paying attention, which is precisely the condition fraud operates in. Every advisory we publish ends with something a person can do this week, and most of those things are free.