The AI training rights thing makes me wonder: presumably, having to grant them permission in the ToS implies that without being given permission, there is at least some situation in which they would not be legally allowed to train an AI on some data? If that is the case, how should a case be legally handled where a user that does not have permission to do so from whoever legally owns said data, posts it and it gets used in the AI training? Presumably if this came up before AI training, all that would be needed to remove the content once the legal owner found out would be some takedown process and the relevant posts would be removed. But, as far am Im aware, you cant just remove something from an AI’s training data and have it “unlearn” that thing as if it had never been included.
To be fair, when people that would otherwise have cause a detriment die before they can do so, one doesnt tend to notice them at all