Who Wrote This?
The IT File
Who Wrote This?
Ideas • AI • Research • Editing • Money • Several Uninvited Variables
Start With the Obvious
One idea. One creator. One finished thing.
That is the clean version.
Then the creator researches.
A tool drafts.
Software corrects.
Earlier ideas influence the new one.
An editor decides what survives.
A reader interprets the finished work differently than its creator intended.
Somebody paid for the software.
Somewhere underneath all of it are influences nobody remembers having.
Assign the Credit
The natural next step is to give everything a percentage.
Percentages make uncertainty look much more official.
Why Are They Doing This?
AI companies have increasingly discussed provenance, watermarking, transparency and disclosure around AI-generated material.
The stated purpose is sensible enough.
People should have information about where material came from and how it was produced.
Anthropic has discussed transparency and research into methods such as watermarking and other approaches to AI provenance.
The difficulty begins when “where did this come from?” turns out not to have one answer.
The current page links directly to Anthropic’s own public material on transparency and responsible AI development.
We May Be Asking Three Different Questions
Authorship
Who contributed ideas, words, research, judgment, editing or execution?
Ownership
Who legally controls the result through contracts, licenses, employment rules or copyright law?
Compensation
Who gets paid? Which may be the author, owner, tool provider, employer, platform, or somebody else entirely.
Does Paying for the Tool Change the Credit?
Argument A: the human paid for the tool. The tool worked for the human. Human credit goes up.
Argument B: the human paid precisely because the tool could perform work the human otherwise would have performed. Human credit goes down.
Both arguments sound plausible.
Which is not helping.
Payment clearly enables creation.
But enabling creation is not necessarily the same thing as creating.
Standard existential uncertainty.
Put Something Else Through the Machine
A song.
A photograph.
A newspaper article.
A recipe inherited from somebody’s grandmother and altered six times.
A film assembled from hundreds of creative decisions.
A scientific paper built on decades of earlier research.
The machine becomes less confident with every specimen.
Which is unfortunate because the machine was designed specifically to create confidence.
Provenance Receipt
Case Status: Solved
And now we know.
...
Wait.
Who wrote this again?
Lanara Tara’s Note
I think the machine is answering the wrong question.
Maybe authorship was never a pie chart.
Somebody starts it. Somebody changes it. Somebody researches it. Somebody rejects half of it. Somebody recognizes something from somewhere they no longer remember.
Then a system asks for one name.
Maybe the strange part isn’t that AI complicates authorship.
Maybe it just makes the complication harder to ignore.
Source Trail
Anthropic has publicly discussed transparency, responsible AI development and technical approaches for identifying or disclosing AI-generated material. Visit Anthropic.
Authorship, ownership and compensation are separate questions. A disclosure system may address one without resolving the others.
Universal Authorship Calculation™ remains entirely fictional. Which unfortunately does not prevent it from looking authoritative.