AI copyright issues are tough to deal with. The training data should ALL be paid for and the original artist or writer should be compensated for the use of their work or it should not be used for training purposes. As far as copyright for AI generated work, another difficult and potentially expensive route is that both the person who created the queries and the AI software company should jointly own the copyright. Some sort of contract with money changing hands needs to happen before anyone can claim copyright or make money from the transaction.
Messy and hard, yup, but it forces companies and individuals to truly take ownership including the financial burden that goes with, especially if money is charged for the resulting art.
As an AI user, I’d like to own the copyright myself, but if I had to share it, it shouldn’t be with some shady software company, but with the original artists. That’s probably extremely hard to keep track of, though 🤷♀️
If you want to own your art, writing music create it all yourself, or in the case of music use a company like Soundraw that only sample music from real musicians who have contributed their original work to the database and are paid fairly. These are NOT large language models or the kind of AI that many fear. https://blog.soundraw.io/post/ethical-ai-in-music
Of course, not all AI is created equal. I'm new to Soundraw, but they seem to have an entirely different approach from 'scavenger' companies like Suno. Only using contributions from willing participants is a good start.
The way Soundraw includes user input could have copyright implications as well. If there's enough user input, even AI productions can be protected by copyright.
Thank you, I follow this music attorney who has a lot of excellent legal advice covering the topic of AI. You might find her worthwhile, if you aren’t already familiar with her work.
AI copyright issues are tough to deal with. The training data should ALL be paid for and the original artist or writer should be compensated for the use of their work or it should not be used for training purposes. As far as copyright for AI generated work, another difficult and potentially expensive route is that both the person who created the queries and the AI software company should jointly own the copyright. Some sort of contract with money changing hands needs to happen before anyone can claim copyright or make money from the transaction.
Messy and hard, yup, but it forces companies and individuals to truly take ownership including the financial burden that goes with, especially if money is charged for the resulting art.
As an AI user, I’d like to own the copyright myself, but if I had to share it, it shouldn’t be with some shady software company, but with the original artists. That’s probably extremely hard to keep track of, though 🤷♀️
You are absolutely right: tighten the existing copyright laws to minimize exploitation.
If you want to own your art, writing music create it all yourself, or in the case of music use a company like Soundraw that only sample music from real musicians who have contributed their original work to the database and are paid fairly. These are NOT large language models or the kind of AI that many fear. https://blog.soundraw.io/post/ethical-ai-in-music
Thank you for this important addition!
Of course, not all AI is created equal. I'm new to Soundraw, but they seem to have an entirely different approach from 'scavenger' companies like Suno. Only using contributions from willing participants is a good start.
The way Soundraw includes user input could have copyright implications as well. If there's enough user input, even AI productions can be protected by copyright.
Thank you, I follow this music attorney who has a lot of excellent legal advice covering the topic of AI. You might find her worthwhile, if you aren’t already familiar with her work.
https://www.youtube.com/@topmusicattorney