|
With the way generative AI works, I really do not consider it to be "createive work". The AI company uses existing data to train the shxt out of their system, and then uses all that existing data to cobble something together. They are creating "new" content alright, but it isn't "creative" or "novel".
If the system doesn't get any useful training data to train its generative model, it doesn't know how to produce any of its outputs. NY Times is suing OpenAI and Microsoft for copyright infringement and profitting off of it. The lawsuit is still gonna take a long time to conclude, but if the verdict is in favour of NY Times, it is gonna become a monumental precedence on how generative AI can continue to evolve. At that point, the cost of producing new generative AI content will almost certainly include the cost to acquire all that training data.
As to the value of the AI-generated content, it is gonna be determined by how much the market is willing to pay, and how greedy the AI companies want to be.
|