Both are true. Brute forcing zips is also faster than brute forcing almost anything else. Other formats use key derivation functions like PBKDF2-SHA1 (hundreds of thousands of iterations of sha1) to slow down the calculation of the key from the password, but PKZIP does not do this. Brute forcing zips can be done at 10 billion passwords per second on a typical GPU, whereas rar/7z/keepass are only a few thousand per second.
Here's an interesting research paper describing both the known plaintext attack and the standard brute force attack https://www.scitepress.org/Papers/2019/73605/73605.pdf
I'm not really sure what to think of this. On one hand, the way I see it, AI deep fakes are essentially a form of defamation, and can harm people by in a way being a false rumour about their sexual life. However, public figures are subject to a much higher standard for defamation, and for a very good reason, else there would be a strong chilling effect on satire, parody, and criticism.
In general I think that deepfakes are only wrong (defamatory) if a reasonable person couldn't easily distinguish them from reality, so obvious fake stuff doesn't count. But for those that are, where is the line drawn for public figures? It is unfortunate that many people can't choose whether to become a public figure, but it is essential to a functional society that freedom of the press and free expression be lenient when it comes to satirical, critical, creative, and even indecent works related to them. But this is of course not absolute.