Who is responsible for work done with AI?
Every project needs someone who can say: I made this decision, and I am accountable for it. AI cannot sign — and a world of unsigned work is a world without trust.
The signature is the foundation of trust
When a doctor signs a diagnosis or an engineer signs a structural calculation, they are not doing paperwork: they are saying ”if this is wrong, I am responsible.” All social trust — in bridges, in medicines, in financial statements — rests on chains of people who sign. AI produces content without signing anything: this is why the output of a model, on its own, is never a document to be trusted. It becomes one when someone makes it their own.
Why AI cannot be held accountable
Being accountable means being able to face consequences: losing one's reputation, one's job, the trust of others. A model possesses none of this; even its developer is accountable for the product in general, not for your individual choice to use it in a given context. Accountability, like water, always flows to the nearest human: if you use AI for a piece of work, you are accountable for that work. Knowing this beforehand changes the way you use it.
Accountable to whom
In impact projects, the chain is concrete: toward the client, who makes decisions based on your numbers; toward the team, which builds on your piece; toward those who live in the territories touched by the project; toward whoever will read the register and trust your Impact Score. Every piece of data declared carelessly is a small loan taken from the trust of all these people. The responsibility is to pay it back with interest.
How to sign work done with AI
Three conditions make a signature honest: knowing how the work came about (which parts the AI proposed, which you decided), having verified what is verifiable, and being able to defend every choice if questioned. If one of the three is missing, you are not signing: you are merely putting a name to something. Simbial's qualification test tests exactly the third condition.
The Impact Controller: responsibility as a profession
On Simbial there is a role built entirely around responsibility: the Impact Controller, who validates the impact data declared by projects. Their signature transforms a declaration into proof, and their independence from the delivering team is guaranteed by the system. It is living proof that in the impact economy, responsibility is not a cost — it is a value that makes you sought after.
The competitive advantage of reliability
In a world flooded with generated content, the ability to stand behind what you deliver becomes rare — and what is rare has value. Those who build a verifiable track record of signed and maintained work — which is the meaning of the Verified, Pro, Expert, Champion scale — possess the professional asset hardest to copy: accumulated trust.
Frequently asked questions
If the AI makes a mistake, whose fault is it?
The fault lies with whoever used the output without verifying it and made it their own. Responsibility always follows the human closest to the decision: this is why work done with AI must be understood, verified, and defensible before it is signed.
What does it take to honestly sign work done with AI?
Knowing how it was generated (what the machine proposed, what you decided), having verified what is verifiable, and being able to defend every choice if questioned.
What is the role of the Impact Controller on Simbial?
Independently validates the impact data declared by projects: their signature transforms declarations into evidence. It is responsibility turned into a profession.
← The full course · Next pill: Ethics in practice: real dilemmas →