To determine whether to run a robot Twitter needs to see millions of accounts a week

Parag Agrawal, Twitter’s chief technology officer, said Thursday that the company challenges millions of accounts every week to determine whether they are operated by humans or completely automated,media reported. The use of fake accounts has been a major concern for technology companies, especially after revelations in 2016 that Russian hackers had used social networks to stir up disinformation during the U.S. presidential election.

Agrawal told the Family Technology Conference that there are all kinds of fake accounts that can be used by bad guys.

Some accounts are fully automated, others are handled by robots, others are run by humans to try to manipulate conversations, and these accounts are attacked and then exploited by bad guys.

In addition to challenging accounts, Agrawal says Twitter closes millions of accounts every month until users see them in timeline or search results.

To determine whether to run a robot Twitter needs to see millions of accounts a week

In addition, he said, companies use technology to monitor the behavior of a large number of accounts and discover patterns used in them. Once Twitter found it couldn’t explain the “abnormality” with “normal use,” the company knew if it should investigate the accounts.

But Agrawal says it’s not easy to tell if a Twitter account is fake because some accounts don’t have avatars and the limited number of tweets they post look like robots but are actually run by humans. Twitter also allows users to use pseudonyms.

But he points out that accounts that don’t look unreal may be the biggest problem, and that “the most dangerous fake accounts don’t actually look fake on the surface.” “