Chapter 17 · AI and society
Copyright and training data
Whether training on copyrighted work is allowed is debated and differs by country and case.
Preparing copyright and ai training data
Chapter 17 · AI and society
Whether training on copyrighted work is allowed is debated and differs by country and case.
Models learn from huge amounts of human work. Who may use that work to train a model is a live debate, and different countries answer it differently.
To train a model, a company gathers a large amount of text, images or code. Making and using those copies is where copyright questions begin.
Read each example and choose where it sits. Then see the factors people weigh. These are teaching groupings, not legal rulings, and your country may differ.
0 of 8 sorted.
Is it settled worldwide that training on copyrighted work is allowed?
Which is generally the easiest case?
Why separate input from output?