We were thinking about this before it was mandatory.
Ethyka comes from a simple conviction: an AI's ethics can't be a PDF nobody applies. It has to be declarable, machine-readable and verifiable.
We've been in AI ethics since 2018.
We started talking about AI ethics in 2018, when almost no one did and it wasn't a trend: it was conviction. We defended it in the media, on stages and in projects when it still sounded like science fiction.
Today the EU AI Act turns into law what was then a concern of a few. Article 5 bans manipulation. We're still in the same place —only now the rest of the world is looking here too.
Four convictions we don't negotiate.
Ethics isn't a PDF, it's a code.
Principles that can't be executed are decoration. That's why the standard is files an agent reads and obeys, and an auditor verifies.
Persuading isn't manipulating.
That one line holds the whole difference between serving a person and using them. Our job is to draw it in writing and defend it with tests.
Open wins.
The rules of machine ethics can't have an owner. Ethyka is CC BY 4.0: anyone can adopt it and improve it without asking permission. That's how lasting standards are born.
Declaring is easy; complying is the work.
Anyone can promise. An agent that keeps its word when someone tries to break it —that's where the value is, and where we stand.
Ethyka is a project by Acuilae Labs.
Ethyka is built by Acuilae Labs, a Spanish company based in Madrid that has spent years working at the frontier between technology, people and ethics.
We talked about this before it was timely.
The standard is open. Start by declaring.
Generate your agent's ethics.md for free, or let's talk about verifying it complies.