The gist of the lawsuit is that Microsoft's use of open source code as input data for it's AI-powered 'Copilot' software authoring tool violates the open source licenses attached to the code. I think we can see pretty clearly how this line of reasoning would apply to other types of AI that use publicly available data such as user photos and images, comments and blog posts, classwork and test responses, and more. Can this line of reasoning work? It's really hard to say; it's a classic problem of composition: no one contribution is responsible for the final outcome, but the final outcome is not possible without all the contributions. There's also an argument by analogy that applies: as individual people, we are influenced by millions of data sources, including private conversations, public acts, and copyright content. Do those sources therefore have a claim on whatever we produce as output? Well, no - but as I've argued in the past, maybe we don't have the claim to originality and ownership over whatever we create that we thought. More: the Verge.
Today: 2 Total: 107 [Direct link] [Share]
