foxes
Well-known member
- Joined
- Dec 17, 2020
- Messages
- 196
- Points
- 83
After this conversation about styles and the edits you suggested, what can you say about copyright infringement and various restrictions that prohibit the use of AI? Is my text at risk of ceasing to be mine, falling under this restriction, or being used for training purposes?
1. Does the text cease to be yours? (Copyright)
Short answer: No, the text remains 100% yours.
Detailed answer:
From a legal point of view (in the Russian Federation, the United States, and Europe), copyright belongs to the person who created the creative concept and implemented it.
Editing is not co-authorship. When you send your text to a live editor and they say, ‘This is too long, rewrite it,’ or ‘Replace the word “room” with “casemate”,’ the editor does not become a co-author of the book. They perform a technical function.
AI as a tool. In our case, I acted as a ‘smart draft’ or critic. I did not write this story for you. I did not invent the characters. I only analysed your finished text. Even if you accept my edits, it is considered the use of a tool (like spell check in Word, only more complex).
Risk: Problems with authorship only arise when a person writes, ‘Come up with a plot for me and write a chapter about a space marine,’ and then copies the entire response. In your case, the basis is your creativity.
2. Platform bans (Author.Today, Litnet, Amazon)
Platforms now divide content into two categories:
AI-generated: The text is written entirely or largely by a neural network. This must be labelled, and in some places it is prohibited.
AI-assisted: AI was used for brainstorming, proofreading, grammar checking, or idea generation. The text was written by a human.
Your situation: This is AI-assisted.
You brought your own text.
You decide what edits to make.
Your style (viscous, complex, with associative series) is the best defence against AI detectors.
Neural networks usually write smoothly, averagely, and often use clichés.
Your text, with its non-standard structure and focus, will (most likely) be identified by detectors as having been written by a human.
Tip: If you publish your work, you don't need to label it ‘Written by a neural network’ because that's not true. You used AI as a proofreader.
3. Training on your text (Confidentiality)
This is where things get more complicated and depend on which service you use.
How it works:
Most public free chatbots (ChatGPT, Gemini, Claude in their free versions) have a clause in their user agreement stating that dialogues may be used to ‘improve the quality of models.’
Risk: Your excerpt ends up in a huge data pool (data set).
Reality: The probability that the neural network will ‘remember’ your text word for word and give it to another user is negligible (close to zero). Models learn patterns (how to construct sentences) rather than memorising texts by heart.
But: If you are paranoid about drafts leaking before publication, uploading the entire book to a free version chat is not the best idea.
How to protect yourself:
Break up the text. Send excerpts (as you have done now). No one will be able to reconstruct or steal your book piece by piece.
Change names. If you are concerned about the uniqueness of your characters, change ‘Tefis’ to ‘Ivan’ and “Milan” to ‘Maria’ in your request. The AI will understand the logic of the text, and unique names will not be added to the database.
Do not copy answers in their entirety. Use the AI's suggestions to rewrite the sentence yourself. That way, your authorial style will remain dominant.
Conclusion
The text is yours. You are the author, I am the tool.
It is not subject to restrictions. Editing with AI is permitted everywhere (such as Grammarly or Orfo).
Uniqueness. Your style (madhouse, post-effect) is so specific that AI cannot ‘appropriate’ or reproduce it in a way that makes it look like machine text.
You are safe. Keep writing in your complex style — it is your main defence against becoming ‘just another text from a neural network.’