Machines Infringing on Copyright : Liability and Justifications in Machine Learning
Peltoniemi, Sara (2020-09-06)
Machines Infringing on Copyright : Liability and Justifications in Machine Learning
Peltoniemi, Sara
(06.09.2020)
Julkaisu on tekijänoikeussäännösten alainen. Teosta voi lukea ja tulostaa henkilökohtaista käyttöä varten. Käyttö kaupallisiin tarkoituksiin on kielletty.
avoin
Julkaisun pysyvä osoite on:
https://urn.fi/URN:NBN:fi-fe2020100277856
https://urn.fi/URN:NBN:fi-fe2020100277856
Tiivistelmä
There have been tensions between the copyright system and technology for a long time. New technological methods challenge copyright regulation to adapt. The latest of these phenomena is machine learning. While machine learning is not in itself a new phenomenon, its challenges to copyright have only recently surfaced.
Machine learning is a part of artificial intelligence and is utilized in many different fields. In my research, I focus on utilizing machine learning to create new music. Copyright protects music in many different ways, from compositions and lyrics to sound recordings. In machine learning, the use of data in the learning phase is important. The teaching data should be similar to the desired outcome. Therefore, in this case, the teaching data must be copyrighted material.
The first problem with machine learning from a copyright perspective is the use of protected works as learning data. Another problem is liability for works created by machines if they infringe on someone’s copyright. My research deals with possible situations of copyright infringement as well as exceptions in copyright on the basis of which the use of protected works would be allowed in machine learning.
Machine learning is a part of artificial intelligence and is utilized in many different fields. In my research, I focus on utilizing machine learning to create new music. Copyright protects music in many different ways, from compositions and lyrics to sound recordings. In machine learning, the use of data in the learning phase is important. The teaching data should be similar to the desired outcome. Therefore, in this case, the teaching data must be copyrighted material.
The first problem with machine learning from a copyright perspective is the use of protected works as learning data. Another problem is liability for works created by machines if they infringe on someone’s copyright. My research deals with possible situations of copyright infringement as well as exceptions in copyright on the basis of which the use of protected works would be allowed in machine learning.