If a "deepfake" suggests someone did something they never did, couldn't that be considered "defamation" (with a couple of other tests)?