Twitter is starting to label bots and automated accounts

The Twitter account NYT_first_said tweets phrases from The New York Instances which have by no means appeared within the newspaper earlier than. Though the Instances itself has disclosed that the account is an bot, not the opinion of an attentive reader, the truth that the account is automated isn’t essentially apparent when its posts present up retweeted in your timeline.

In order that customers not have to guess which accounts are automated bots and that are run by people at keyboards, Twitter is testing the addition of a label that may seem alongside bot account names indicating the account is automated and offering a hyperlink to the Twitter account of the one that runs it.

The brand new function comes after the corporate’s analysis revealed that customers are a bit confused concerning the function of automated posting on the platform, says Twitter product administration director B Astrella. It’s a part of a normal push to assist folks perceive particular sorts of accounts that prospects have discovered complicated, together with bots, he says. Astrella indicated that Twitter may additionally quickly supply particular labeling for memorial accounts which can be preserved after an individual’s demise.

The objective isn’t a lot to ferret out malicious bots attempting to mislead customers, he says, however to level out how “good” bots work. Astrella says such good bots embrace automated accounts that submit realtime earthquake data, share ASCII art, share data about what’s in the newspaper, or simply remind folks to stay hydrated.


Whereas the system is presently being examined with a couple of hundred accounts, Astrella expects that it’ll in the end be required of all bots on the platform, which as of final yr are already supposed to disclose of their profiles that they’re bots and point out who made them. The brand new system will standardize that label reasonably than making folks parse it out of the account description area and will publicize that accounts are bots the place their tweets seem in consumer timelines, not simply when folks click on via to their profiles.

[Image: courtesy of Twitter]

Good bots and unhealthy bots

There isn’t a hard-and-fast definition of what counts as a bot: Astrella stated it’s typically an account that posts greater than half its posts robotically, however that different accounts may require the label in the event that they, say, periodically submit bursts of automated materials.

Not all bots are good. Automated accounts have been used previously to share misinformation and propaganda on the platform: A research final yr, as an example, discovered that almost half of customers then sharing details about the coronavirus pandemic were bots. Numerous analysis has over time suggested that between 5% and 15% of accounts on the positioning are pretend and Twitter has beforehand taken down swaths of bots and different pretend accounts used by governments for political manipulation.

The corporate says it has guidelines to be sure automated accounts don’t attempt to manipulate conversations on the social community or annoy customers. As an example, bots aren’t allowed to ship bulk, unsolicited direct messages, and they’re not allowed to robotically like content material on the platform. The corporate has invested in rooting out such “inauthentic behavior” on the platform, which incorporates customers posting industrial spam and bots making an attempt to drive conversations via tweeting or popularizing posts. That can proceed, Estrella says.

Bots that presently attempt to masquerade as non-automated accounts could attempt to violate the brand new rule, even perhaps making the most of the coverage to mislead customers who assume all bots can be labeled as such. Astrella says he doesn’t presently see a variety of bots masquerading as actual customers.

“Within the locations the place I’ve seen of us not disclosing, it’s principally simply because they don’t know that they’re supposed to,” he says.

Astrella says the suggestions he’s heard from builders has typically been optimistic, particularly because it provides them a extra official approach to take credit score for his or her bots.


“This provides their possession of the bot slightly bit larger standing on the profile,” he says.