Countries seeking to become world leaders in AI development face a copyright dilemma: should it be copyright infringement to train AI models on large datasets consisting of copyrighted works? If so, what if AI other countries allow it and gain a competitive advantage in AI technology?
Right now, the UK government is embroiled in a contentious debate with many artists speaking out against the initial proposal of an “opt out” system by which authors would have to affirmatively opt out of the use of their works. Authors argue that turns copyright law on its head by putting the onus on them. Instead, it should be an opt-in system: no unlicensed uses are permitted.