On: America.gov

[davesque]

I'm a bit concerned that a government chatbot like this would include a user-feedback signal like thumbs up/down. I think this could be gamed by bad actors to promote/demote specific types of content in the training data, assuming they use those signals in some sort of post training. Hopefully they're accounting for that.

[tancop → davesque]

You can combine it with a topic classifier and get the satisfaction rate for each topic. That's pretty useful for finding problem areas.

[vuldin → davesque]

... unless it doesn't really even do anything and is just a placebo.