Nina O'Sullivan has written an article for PDP journals summarising the current state of case law and policy developments around the use of copyright works to train GenAI models and the extent to which IP protection is available where GenAI models are used to create a work.
An excerpt is below:
Three years following ChatGPT's public launch, generative Al's explosive growth has created profound challenges for intellectual property rights owners. Legislators worldwide are grappling with how best to meet competing imperatives: continuing to provide robust IP protection, whilst encouraging AI investment and national competitiveness.
Meanwhile, the courts have also been inundated with GenAI copyright lawsuits, with over 60 cases having been launched worldwide. Whilst the battlelines continue to be drawn, the first settlements are also being signed, suggesting that a licensing equilibrium will ultimately be reached between AI model developers and rights holders. The risk though of such an outcome (as seen before with music streaming) is that this may come at the expense of individual creators....
This excerpt is adapted from an article originally published in the AI Compliance & Risk Journal, a PDP Journals publication.