Recently, the social network giant Facebook released its first transparency report on the actions of moderation that runs on the social network. The document reveals truly impressive numbers: in the first quarter of 2019, the company eliminated 583 million fake accounts in the service, in addition to 837 million spam publications.
OMG! Facebook Deleted 583 Million Fake Accounts
The social network giant Facebook released its first transparency report on the actions of moderation that runs on the social network. The document reveals truly impressive numbers: in the first quarter of 2019, the company eliminated 583 million fake accounts in the service, in addition to 837 million spam publications.
The first amount impresses even more if we take into account that in the period, Facebook registered 2.2 billion active monthly users (those who access the social network at least once a month). This means that there is more or less a fake profile for every four legitimate accounts.
It’s an old problem on Facebook, so much of the legitimate users have probably been added by fake profiles. It is not difficult to identify them. Usually, fake accounts use photos that draw a lot of attention, do not have many publications and interact less.
Despite the gigantic number of fake accounts, Facebook points out that most of them were deleted minutes after they were created, preventing them from reaching a significant number of legitimate users.
This filtering is done with the aid of artificial intelligence. Even so, the company estimates that from 3% to 4% of the active accounts in the first quarter corresponded to false profiles.
It is also thanks to artificial intelligence that Facebook has managed to tackle the problem of spam. According to the company, almost 100% of the 837 million such publications have been automatically detected and deleted, even before users have a chance to report them.
In the report, Facebook points out that 21 million publications with scenes of nudity or sexual activity were eliminated, 96% automatically, without users having to report the irregularity. 3.5 million publications with violence were erased or warned, 86% of them by the system itself.
These are amazing numbers, but Facebook itself acknowledges that this is a work in progress. Filtering systems already work well but are not effective in identifying hate speech, for example. 2.5 million publications of the type were eliminated in the period, but only 38% automatically.
The social network giant Facebook also explains that artificial intelligence algorithms need a lot of data to be trained, which takes time. In addition, there is the question of the reaction: the company points out that those responsible for irregular publications are constantly changing strategy to circumvent social network controls.
So, what do you think about this? Simply share all your views and thoughts in the comment section below.