Every time people go to Twitter to discuss Charlie Sheen’s latest pronouncement, they pick up new followers with mysterious handles like @kysuvygix and @upysoxy. Those aren’t Basque fans of Two and a Half Men. They’re Twitter bots—programs that follow real users based on popular keywords like #winning.

Do they pose a risk to you? Not as long as you ignore them. The people behind them are hoping to appeal to your narcissism. If you’re someone who reflexively follows anyone who follows you, you’ll be inviting spam messages until you unfollow or block the offending accounts. And if you like to read the notification emails from new followers, you’ll often find, in the case of the bots, links to scam pages.

One website that sells Twitter bot software promises to acquire 1,500 to 3,500 followers per week. The software lets spammers control multiple accounts, mask their IP address via proxies, and find followers by location or time zone for targeted hits.

Twitter hunts down and kills whatever bots it can. As has long been the case with email spam, though, the company is in an arms race. Part of the reason the bots stay ahead is Twitter’s policy of not monitoring content, which would be a fairly straightforward way of spotting spam. Instead, Twitter’s algorithms search for behavioral cues, kind of like the Blade Runner test. Does a new account send lots of @replies to people it doesn’t follow? Are its messages mostly the same? Such signatures can trigger an automatic shutdown.

“But not all bots send spam,” says Del Harvey, Twitter’s head of trust and safety. “And not all spammers are bots. There are behaviors that look spammy but instead indicate a person in an emergency.” For example, someone trying to find out if a wildfire is coming toward their house might send the same @ message to several people they don’t know, much as a bot would.

“That’s a false positive I really want to minimize,” Harvey says.

Thus, Twitter relies primarily on human intuition—on people reporting spam via the user pages of the bogus accounts or by direct messages to @spam. Those actions alert Harvey’s nine-person antispam team to investigate the accounts and, if need be, retire the offending bot.

So Twitter is yet another battlefield in the war between humans and machines. And the humans need all the help they can get.