What a shock, bots are likely driving much of Twitter's 'reopen America' conversation

Researchers at Carnegie Mellon University say that nearly half of all accounts tweeting "reopen America" are likely bots. The sock puppet accounts are flooding Twitter at two times the level seen in previous domestic and geopolitical events. Some of the bot accounts are also promoting conspiracy theories of the coronavirus being linked to 5G technology, an idea that has been repeatedly debunked.

Coordinated bot activity on Twitter is often a means with which to advance a political agenda and increase polarization between groups. By pushing the message to "reopen America" from puppet accounts that are indistinguishable from real people, elected officials may get the wrong idea that there's increasing support for lifting shelter-in-place orders. Some states have already begun reopening businesses, but it's yet to be seen what effect doing so will have on infection numbers.

Detecting a bot — CMU's researchers say there are several signals they look at to determine whether or not an account is a bot, including follower account, frequency of tweeting, and the number of @-mentions the account receives. Tweet frequency is often a clear indicator, as bots typically tweet way more than an actual human would. It's also common to see the exact same or similar messages being tweeted from different accounts in rapid succession.

Twitter is a particularly attractive target for bots because it's one of the few places where government offices and public citizens diverge in an open forum. But that makes it a juicy attack vector, with political messages sometimes rising to the top on the back of a swarm of fake accounts that appear to be coming from thousands of people when in fact it's just a few people working in coordination.

Unique dangers of this activity — Coordinated campaigns to reopen America are particularly insidious as doing so too early could have fatal consequences, especially for people who are immunocompromised. Trump earlier today called on governors to allow places of worship to open this weekend, even as the death toll from coronavirus nears 100,000 and experts warn of a potential second wave of infections in areas with easing restrictions. More than 25,000 new cases were reported on Thursday.

At the same time, many people are reasonably anxious about lost income as time drags on and the unemployment rate remains at record highs. Bots could have a serious influence on these groups.

Even Elon Musk, a master at mobilizing his legions of fans to attack people on Twitter, has implored the company's CEO Jack Dorsey in the past to do more to differentiate between real and fake users. "Basically, how do you tell if the feedback is real or someone trying to manipulate the system, or probably real, or probably trying to manipulate the system?"

Twitter consistently says it's working to detect spammy behavior and bot manipulation, but the accounts often find new ways to evade the company's grasp. Facebook, too, continues to battle coordinated activity, and releases monthly reports updating the public on all its new account deletions.

CMU's researchers say that you should take the typical precautions to prevent yourself from being influenced by fake accounts. Closely examine accounts for indications like high tweet frequency or sharing of links from unfamiliar news sites.