- There is a pressing need for clear regulations around artificial intelligence’s (AI’s) use of copyrighted material, especially regarding training data.
In October last year, Microsoft CEO Satya Nadella made his contribution to the ever-growing body of commentary on the conflict between artificial intelligence (AI) technology and intellectual property (IP) infringement. As the defendant in several copyright infringement proceedings worldwide, it probably shouldn’t come as a surprise to learn that Microsoft considers that making use of copyright works for the purposes of training machine learning AI models should not amount to infringement.
Nadella questioned why a machine learning model is any different from someone reading a textbook and using the information they take in to produce new content. While his provocative conclusion, that reading a copyright work and then creating ‘new knowledge’ could be infringement, falls wide of the mark, he is certainly not alone in demanding clarity on how copyright laws
