If humans have a bot-like behavior, it's okay to mark them as bot. If a human is only posting to promote products/astroturf, who cares if it's misclassified, it doesn't add anything to the discourse. IMHO, that's good riddance.
And in my solution, at the end the instance owner takes action, it's not like there is no human recourse.
Because there is no karma system on lemmy (thank goodness, I'm against karma), you can easily create a 1000s of bots which will upvote your post and bring it to front page.
The solution is not some custom anti-abuse system which can be game. (Stuff like "you can't vote because of the age of your account", ...) IMHO, the solution is bot detection. Since everything is public an an instance, somebody at some point will start scraping instance to detect bot behavior and inform instance owner. It will come with maturity.