Artificial intelligence (AI) has been a hot topic this year. There have been debates around AI regulation, the intellectual property (IP) rights of AI, and about copyright protection in the age of generative AI. We did several stories on these topics and have summed up the most important points from them here: Whose data is it anyway?: We attended the UNESCO conference on Generative AI and intellectual property protection on May 25. One of the biggest questions that emerged from this discussion was which data can be protected as IP. Besides the obvious concern surrounding copyrighted data used in training AI models, the question is: how do we categorize user prompts and even the output emerging from entering these prompts (and refining them to get the right output)? Should this fall under the purview of human authorship? This question of data ownership becomes even more complicated when we consider augmented data and synthetic data, both of which aren't direct replications of pre-existing human expression but use it in one way or another to create training data for AI models. Can companies classify the use of copyright data as “inspiration”: In an interview last month with the creator of the AI inventor DABUS, Dr Stephen Thaler, we asked him about the ethics of using copyrighted material to train AI. Thaler claimed that doing so is the equivalent of an artist taking inspiration from the Mona Lisa. But copyright activist Neil Turkewitz disagrees. He believes that corporate ingestion of art for the…
