Training AI Models

What does “Training” of a model mean?

“Training” is the process through which an AI system analyzes large volumes of data in order to learn patterns and structures.

In music, this data may include:

  • recordings
  • musical scores
  • lyrics
  • metadata
     

What are datasets?

Datasets are collections of data used for the training of a model. The quality, origin and legality of this data are of critical importance.

When datasets include protected works, issues of lawful use arise.

 

Licensing, Consent and Compensation

The use of protected works for AI training may demand license from royalties holders.

Major issues include:

·       explicit consent

·       fair remuneration

·       the determination of terms of use

Ensuring these conditions is a key pillar for the sustainable development of the technology.

 

Transparency and Traceability

Transparency regarding :

o   which works are used

o   on what way

o   for what purpose

The creation of a relationship of trust between technology providers and the creative community is considered a necessity.

Traceability contributes to the detection of possible violations and the securing of fair remuneration.

 

The role of Collective Management Organizations

Collective Management Organizations can play a significant role:

o   in negotiating licensing

o   in collecting and distributing royalties

o   in shaping best practices

o   in representing creators at national and international level

 

 In a rapidly evolving technological environment, the protection of creation remains a fundamental principle.

 AUTODIA supports innovation, while respecting the rights of creators, aiming for a balanced and fair framework for the future of music and Artificial Intelligence.

Share
Autodia Collective Management Organization of Music Authors & Rightholders