I haven't read the book itself, but it's obvious that 'immutable audit logs' don't prevent privacy violations. They would simply make these privacy violations d

I haven't read the book itself, but it's obvious that 'immutable audit logs' don't prevent privacy violations. They would simply make these privacy violations detectable. Enforcement of privacy violations would still be up to the rest of the oversight structure (whether that's institutional, external audits, journalism, special interest groups, or a combination thereof). And either way, technical measures that would be insufficient for PRISM may very well be sufficient for most of the rest of what government does. As a final note I think that sometimes we tech types get way too wrapped around preventing outcomes which are already so implausible as to be merely theoretical, and use that to oppose measures which, while imperfect, would still be useful if applied. It's "The Perfect is the enemy of the good" that we see in software project management, applied to the government at large.