> And training on all human-made media needs to be banned also. You're right. Companies would need to implement draconian "content attribution systems" to scan

> And training on all human-made media needs to be banned also. You're right. Companies would need to implement draconian "content attribution systems" to scan all creative works to ensure no training data was derived from copyrighted sources. This could result in a "chilling effect" where creators self-censor for fear of their original works being flagged as infringing simply due to incidental similarities. Because now we have to scan every human work, as it might have been generated with AI. It opens up the prospect of expensive litigation over subconscious inspiration or vague resemblances to prior works. Taken to the extreme, it could hamper creativity and innovation by making any new work legally risky to publish if AI analysis deems it too "similar" to something else.