People say — AI wrote that.
No. I wrote it. AI was the instrument. The judgment, the framing, the knowing what to ask and what the answer means — that was mine. Using AI did not reduce the effort. It relocated it. Less on execution. More on thinking. More on the part that cannot be replicated.
The assumption embedded in “AI wrote it” is that the effort disappeared. It did not. And that assumption is not innocent. It is the mechanism by which the relationship between a professional and their work gets erased — before a dispute begins, before a claim is filed, before any legal framework has a chance to activate.
Which brings the problem into focus. The IP debate is not failing because the law is slow. It is failing because it is protecting the wrong thing.
Copyright covers output — but only in the hands of someone with standing to claim it. You cannot file protection for Nat King Cole’s catalogue because the law already recognises that the relationship between you and that work does not exist. The law already knows relationship matters. It just activates too late, applies too narrowly, and costs too much for the people who need it most.
In the United States, AI-generated output is currently not protectable. The human prompt that produced it is. But the protection does not cleanly follow the prompt into the output. The most skilled prompter in the room — the professional with decades of domain knowledge embedded in every instruction — produces work that is legally unowned. The relationship between the professional and the output is real, earned, and irreducible. The law has no mechanism to recognise it.
We do not have a name for what actually needs protecting. For this discussion, I will call it relational authority — the irreducible connection between a person and their work that technology can replicate in form but never in meaning.
In 1991, Natalie Cole recorded a duet with her late father using archival footage and studio technology. Anyone can replicate that production technically. Only Natalie could make it resonate — she was his daughter, and the grief was real. Only the Cole family can commercialise it legitimately.
Three completely different questions. Who can make it. Who can make it mean something. Who can profit from it. One legal framework — and it addresses the third, imperfectly, while the first two go unnamed.
The resonance question is not about copyright. It is about relational authority. The technical access is irrelevant to it. The commercial right is downstream of it. Natalie Cole’s version of Unforgettable is not legally protected from imitation. It is humanly irreplaceable. Those are not the same thing. Conflating them is how this debate stays permanently unresolved.
Most people think the AI and intellectual property debate needs one answer. It needs at least three — accountability, provenance, and trust. They do not operate the same way. They do not stack the same way. And they are not available to everyone in equal measure.
Accountability is institutional. Someone put your name on the brief. Provenance is evidential. There is a record of your involvement over time. Trust is relational. People in the room know what you built and believe your claim without requiring the paper.
The most exposed are those who can access only one. A junior professional. Someone earlier in their career with a thinner paper trail. Someone whose relational authority is real and whose ability to defend it is not. The debate has largely been conducted by those who have all three — about a problem most acute for those who have one.
These three defences are not separate strategies. They are what a defensible position produces.
When a professional documents their methodology, records their reasoning, and signs their name to a conclusion reached with AI assistance — they are not performing compliance. They are making the relationship between themselves and their work visible and evidential before a dispute forces the question. They are answering “in what capacity” before anyone asks.
In the AI age, ownership of an idea belongs to the professional who can demonstrate the judgment behind it. Not the output. The judgment. The AI produced the output. The professional produced the judgment that instructed, shaped, selected, and validated it. That is the relationship. That is what needs protecting.
The defensible position is how you protect it.
The law will catch up. It always does, badly, after the damage is done. What matters now is whether professionals — particularly those earlier in their careers, with thinner paper trails and fewer decades of trust behind them — can name what they actually need to protect before someone else defines it for them.
Accountability, provenance, and trust are not bureaucratic caution. They are how professionals in the AI age assert ownership — not of the output, but of the judgment behind it.
Own the judgment. You own the idea.
We are still early enough that the factors we have not yet named will matter before this is resolved.
That is an opening, not a comfort.
This story above is a well written piece by David Wong MRICS Planning and Development
