Three sentences went up on the front page of covenant.to today, and I want to explain why they’re there, because they’re doing more load-bearing work than their length suggests. The first is about rights, not technology: every person has a fundamental right to use technology for any legal purpose, and “legal” is a question for We the People, decided through our legislative and constitutional institutions — not by whoever happens to own the platform you’re using. AIGCSEP doesn’t get to define what’s allowed. It exists to make sure whatever society decides is allowed actually happens, accountably, at scale.
The second and third are about why any of this exists at all. Human beings were given minds to build, and we’re not going to stop — we never have, and we shouldn’t start now. The only real question has ever been whether we build the accountability structures alongside the capability, or after the fact, when it’s too late for them to matter. And underneath all of it: AI is two things, full stop — a human interface to compute, and a converter of human intent into action and artifact. Every design choice in AIGCSEP, the Vault, the Badge, the three planes, all of it, follows from that one definition. Understand AI as those two things and nothing else, and the protocol stops looking like a list of features and starts looking like the only sane way to build this.
None of this is new — it was always implicit in the RFC, assumed rather than stated. It’s explicit now, where anyone can hold me to it. If you read draft-howlett-aigcsep-00 and think the mechanism doesn’t follow from these principles, or think the principles themselves are wrong, I want to hear it. The draft is open for review at datatracker.ietf.org/doc/draft-howlett-aigcsep/.
— J.P. Howlett
Discussion
Comments aren’t wired up here yet — they’re coming. For now, if this piece sparked a thought, the fastest way to reach me is through the About page.