#MerzSays: OPT-OUT or OPT-IN for AI?
The Situation
WIRED explores a very important topic: to apply AI we should train it. To train it, we shall collect a bigger Dataset (the bigger, the better for generative capabilities — Merzmensch). But the actual datasets are full of contents whose creators are not aware of (yet probably don’t agree with) use it for AI training.
What to do? A possibility for opt-out would be important: the creator has right to explain his/her work were not allowed to be used in this way (and to give opt-out from such dataset collections).
Merzmensch opinion.
We need to raise awareness about dataset contents. And also bring new way of thinking. I, for my part, would love to give my OPT-IN for all training datasets to use all my texts, photos, images, videos I did. Yes, I spend countless days and nights writing and creating, but I want to contribute to the future. I want that my contents belong to DNA of next generations of Generative AI.
Without Question: Everybody should have right to opt-out. But by opt-outing one should be aware that his/her work won’t be relevant - or known - by AI-driven systems of the future, which will be more and more entering our science, society and culture. (Proper) AI models never plagiarize, they use Training Dataset to learn how to write or to draw.
But our entire human culture consists of references and deviations.
