sourch

Facebook is still struggling to curb the spread of terrorism, violence, violent content and spam from their popular platform.

Facebook posted their first-quarter Community Standards Enforcement Report this year, in the last two quarters (the fourth quarter of 2017 and the first quarter of 2018), the social site has disabled about 130 billion fakes or fake accounts, most of which were intended to spread spam or illegal activities such as scams .

The social giant's site also said, in the last two quarters, in the last six months, millions of violations of the policy occurred. Including abuse, graphic violence, adult nudity and sexual activity, terrorist campaigning, spam and fake accounts.

Facebook has acknowledged that their artificial intelligence detection technology is not yet sufficiently efficient - especially in identifying hated data. In this case, it is necessary to check through Facebook moderators.

Guy Rosen, Vice President, Product Management, Facebook, said in a statement: "It is very important that progress needs much more and we will probably change our method, through which we will learn more about importance and work. We have a lot of work to do to prevent posting of policy.

The company had earlier said that they are planning to hire more than a thousand new moderators to keep Facebook safe for everyone.

In the first quarter of this year, Facebook has moderated 2.5 million posts for violation of rules, and only 38 percent of those posts were detected by automated technology, failing to counter counter-attacking comments or signatures in the context of offensive comments.

In addition, Facebook has deleted 34 million violent content in the first quarter and warned that from the last quarter of the year 2017 till the first quarter of this year, it has increased by 134 percent. Of this, about 86 percent of the violence content has been detected by Facebook, before users complain.

Facebook's Vice President Guy Rosen said, "Faced with international disagreements, such as the recent Syrian incident, Facebook's growing violent content can be the reason behind. It is often seen that when a controversial situation arises in the world, the matter comes to a great extent on Facebook.

Facebook has been severely criticized for the government and human rights organizations of different countries, due to rumors, hate content, terrorism, increasing use of the site in conflict work and promoting issues such as murder and suicide through the platform.

Facebook uses both software and moderator to delete disobedient writing, pictures and videos.

Guy Rosen said, "Techniques such as artificial intelligence are not yet good at the point of bad content because the subject matter of the sentence is important. Artificial intelligence technology requires a large amount of trained data, which can understand the meaningful behavior of different behaviors, because it is often lacking in less commonly used languages. '

However, the efficiency of the artificial intelligence technology of Facebook increased the identity of the fake account automatically before the user received the complaint. In the last two quarters (fourth quarter of 2017 and first quarter of 2018), 98.5 percent of accounts accounted for 100 percent of fake accounts associated with spam, and Facebook's artificial intelligence tools.

In the first quarter (January-March) of this year, Facebook has blocked 58 million 30 million fake accounts, within a few minutes of registration of accounts.

More than a quarter of people in the world use Facebook, Facebook's monthly users are 200 million Under pressure from the governments of several countries, Facebook is also increasing their number of moderators. By the end of 2018, the number of Facebook moderators is expected to reach 20,000.

Take a look, from October 017 to March 2018, Facebook has disabled the content of 6 months.

Violence Issues

• October-December 2017: 1.2 million content (72 percent were identified in auto-technology).

• January-March 2016: 34 million content (86 percent were identified in auto-technology).

Hate Speech

• October-December, 2017: 16 million content (24 percent identified in auto-technology).

• January-March 2016: 2.5 million content (38 percent were identified in auto-technology).

Porn

• Oct-Dec 2017: Twenty million 10 million content (94 percent identified in auto technology)

• January-March 2016: 20 million content (96 percent were identified in auto-technology).

Terrorist campaign

• Oct-Dec 2017: 11 million content (97 percent of which were identified in automatic technology).

• January-March 2016: 1 million content (99.5 percent was identified in auto-technology).

Spam

• October-December, 2017: 72 million and 70 million content (99.8 percent was identified in auto-technology).

• January-March 2016: 83.6 million content (99.7 percent was identified in auto-technology).

Fake account

• October-December 2017: 694 million (99 percent were identified in auto-technology).

• January-March 2015: 58 million 30 million (99 percent were identified in automatic technology