As you know, last week,
a whistleblower claimed that high-level plans are afoot to give AI developers some sort of copyright “carve-out”, allowing them to use Australian music, journalism, literature and art to train their commercial products without consent or compensation. The whistleblower also claimed that “billions of dollars of data centre investment” was on the table, with “hundreds of millions of dollars annually for a creative fund” as a sweetener.
If true, this abrupt U-turn on safeguarding against text and data mining would end careers, diminish Australian culture, and exacerbate some of the most perilous crises facing this continent.
What if they simply said that any works containing ai material, where the AI used stolen copyrighted material to train, would not be allowed to be copyrighted and, no royalties allowed, and is immediately public domain works.
And if you want to use AI to make copyrighted works, you must provide evidence of the training material used by the AI, and that it is all above board.