> making it possible to generate . . . comments at scale, drowning out the voices of real citizens in a tidal wave of fake ones. We need to address *that* problem instead of framing this as a problem with "AI" (which, as I have noted before, this current crop of machine learning algorithms should *not* be conflated with). This is very directly related to topics like electronic voting, because it is yet another way the voice of the people can be heard. So whatever process validates public comments should work *regardless* of their content; it should even be effective for simple yes/no polls. > These companies then stole people’s names and email addresses from old files and from hacked data dumps and attached them to 8.5 million public comments and half a million letters to members of Congress. That is a serious amount of identity theft at a minimum, and quite possibly a massive human rights violation. It goes so directly to subverting democracy that it might rise to the level of [treason](https://en.wikipedia.org/wiki/Treason_laws_in_the_United_States). Officers of these companies should be charged with capital crimes. > The danger . . . is that public commentary will be completely discredited. In many corners of the Internet, it already is. Your blog itself is an interesting case study, Bruce, because it still manages to curate mostly quality comments despite allowing mostly anonymous posting. I wonder what mechanisms you use and could stand to improve upon in order to keep other forms of public commenting safe from manipulation at scale. @Chris Dotson >Trusted digital identities issued by the government would solve a lot of problems, although it would probably cause a few too. Well, for starters, they should simply verify the identifying info they're already getting! I can't go to a podunk web site and set up an account without my email address being checked, yet the federal/state government isn't doing that basic of a challenge-response test for the submissions it receives? Makes me wonder if there isn't corruption at a deeper level that needs to be addressed. @Tatütata >IMO, a pretty useless facility, which on this blog and elsewhere is mostly used by s{p,c}ammers While that is true in general, I happen to use it to verify that my identity is linked to my comments here. A copy of most "public commentary" I make is on my blog, and I can give a specific URL to verify that I've actually said what this blog's comment section says I said. That's another solution to the problem being discussed, making fakes a pointless endeavor regardless of how sophisticated their generation may be. a362fd4acac3973e98a4a48fa1a32ef66a92f72b1a460f68cb974d8a9b9031f9