AI и тест за авторско право

Отговорите и обясненията се появяват, докато отговаряте на теста.

1. Какво е заявила Службата за авторско право на САЩ за произведения, генерирани изцяло от AI без човешки творчески контрол?
   • They belong to whichever user paid the most for access
   • They can't be copyrighted because human authorship is required
   • They are owned automatically by the AI system that generated them
   • They receive copyright protection for twice the normal term
2. В Thaler v. Perlmutter, какво реши съдът относно изображение, посочващо AI система като негов единствен автор?
   • The image entered the public domain after just one year
   • The AI was granted copyright protection as a legal person
   • Registration was denied because copyright needs a human author
   • Copyright went to the company that manufactured the GPU
3. В законодателството на САЩ кой фактор за честна употреба разглежда дали дадена употреба е „трансформираща“?
   • The purpose and character of the use
   • The nature of the copyrighted work
   • The effect on the potential market
   • The amount of the work used
4. Защо Гилдията на авторите срещу Google често се цитира в дебати относно обучението на AI върху защитени с авторски права книги?
   • It banned all use of copyrighted books in software
   • Google was ordered to destroy every single book it had ever scanned
   • It gave authors ownership of all search engine results
   • Google’s book scanning and limited snippet display for search were fair use on that record
5. Съгласно изключението на ЕС за извличане на информация за текст и данни (TDM) в Директивата за авторското право от 2019 г., какво могат да направят притежателите на права?
   • Demand deletion of any trained model anywhere in the world
   • Expressly reserve their rights to exclude uses from Article 4’s text and data mining exception
   • Block all academic research that uses their works
   • Nothing, since all text and data mining is forbidden
6. In a copyright dispute about generated text, why can near-verbatim output examples matter?
   • The number of users who subscribe to the paid service
   • The color scheme and fonts of the chatbot interface
   • Outputs reproducing near-verbatim passages from training works
   • The model's total electricity consumption during training
7. A tool permits commercial use of its outputs. What does that permission establish about third-party rights?
   • It does not necessarily resolve rights held by other people
   • It proves every training work was licensed
   • It guarantees copyright registration
   • It overrides local copyright law
8. Why should a publisher assess copyrightability and infringement as separate questions?
   • The questions always have the same answer
   • Infringement only concerns software licenses
   • A service contract determines both questions everywhere
   • Human authorship may affect protection, while use may still implicate someone else’s rights