hii

| On
August 20, 2021

Facebook reveals Community Standards Enforcement Report, Q2 2021

Share

Facebook revealed Community Standards Enforcement Report, Q2 2021. This report contains data on how we implemented our rules from April to June. This is our tenth report, and the following are some of our long-term trends:

Three-quarters of the time since we began reporting on hate speech, it has decreased in prevalence. Improved hate speech detection and News Feed ranking changes are to blame for this.

  • Since we began reporting hate speech content on Facebook and Instagram, removals have increased by over 15X.
  • When it comes to Facebook, we have a proactive rate (the percentage of content on which we have taken action before a user has reported it to us) of over 90% in 12 policy areas and 9 of 11 on Instagram.
  • In addition, we’ve added new metrics on appeals, restores, and prevalence to Facebook and Instagram.

Besides our Community Standards Enforcement Report, this quarter, we’re also sharing:

Updates on our efforts to promote reliable information about COVID-19 and vaccines, as well as to reduce harmful misinformation.

If the data shows areas where we need to improve, we’re committed to sharing it to be held accountable for our progress.To kick off a new series of reports, we’re releasing the top 20 most viewed domains, links, Pages, and posts in the United States. Transparency Center makes these reports available, and we’ll include them in future quarterly Community Standards Enforcement Reports.

Promoting Reliable Information and Reducing Harmful Misinformation About COVID-19

Our commitment to helping people find reliable information, including information on the COVID-19 vaccine, will not change. To this day, we’re removing harmful COVID-19 misinformation and banning ads that try to profit from the pandemic’s spread. From the beginning of the pandemic to the end of June:

  • Over 20 million pieces of content were removed from Facebook and Instagram for violating our COVID-19-related misinformation policies.
  • For repeatedly violating our rules against spreading COVID-19 and vaccine misinformation, we’ve removed over 3,000 accounts, pages, and groups.
  • As a result of our collaboration with 80 fact-checking organizations in over 60 languages, we displayed warnings on more than 190 million pieces of COVID-related content on Facebook that our third-party fact-checking partners rated as false, partly false, altered, or missing context As a result, we add a prominent warningbefore a piece of content is rated with one of these ratings, a prominent warning is added before sharing, and the post is pushed down in the feed.

Using our COVID-19 Information Center, educational pop-ups on Facebook and Instagram, and our vaccine finder tool, we’ve helped 4 million people in the United States gain access to vaccines.

According to public health research, vaccination rates are higher among those who see others in their community getting vaccinated. Vaccines are widely available in countries where the majority of people have access to them

  • Globally, more than 18 million people have used anti-vaccine Facebook profile frames since the outbreak started.
  • Facebook users around the world have seen more than 25% of the UNICEF COVID-19 vaccine profile frames.
  • The COVID-19 vaccine sticker has been used by over 29% of Instagram users worldwide.
  • 7.6 million people have used the Instagram COVID-19 vaccine stickers worldwide.

According to a new study, vaccine hesitancy has dropped by 50% among Facebook users in the United States. We’ve also seen a rise in vaccine acceptance around the world. So, for example, our COVID-19 Trends and Impact Survey data, which we conduct in partnership with Carnegie-Mellon and the University of Maryland, has shown that vaccine acceptance has increased by 35 percent in France, 25 percent in Indonesia, and twenty-five percent in Nigeria, since the beginning of this year.

Community Standards Enforcement Report Highlights

Progress on Hate Speech

For the third quarter in a row, hate speech on Facebook has decreased. A total of 5 views per 10,000 views was recorded in Q2, down from 5 to 6 views per 10,000 views in Q1.

During Q2, we removed 31.5 million pieces of hate speech from Facebook, compared to 25.2 million in Q1; and 9.8 million pieces of hate speech from Instagram, an increase from 6.3 million during Q1. This is because we are constantly improving our proactive detection methods. Facebook and Instagram hate speech violations can now be detected by our investments in artificial intelligence (AI). The Reinforcement Integrity Optimizer (RIO), for example, allows our AI models to detect hate speech using real-world data and to improve over time. This technology allows us to enforce our policies across billions of users in multiple languages and across billions of users worldwide using this technology.

Updates in Child Safety

Making sure our apps are safe for kids to play on is essential. Prior to this, we reported on child nudity and sexual exploitation of children as key metrics. More data and two new reporting categories under the broader topic of child endangerment have been added to our latest report. As well as sexual abuse, there is also nakedness and physical exploitation.

We changed this to give child safety experts, academics, and the general public a more detailed, transparent overview of our efforts in this space.

  • Child nudity and physical abuse content we took action on:
  • On Facebook: 2.3 million with a proactive rate over 97%
  • On Instagram: 458,000 with a proactive rate of over 95%
  • Child sexual exploitation content we took action on:
  • On Facebook: 25.7 million with a proactive rate of over 99%
  • On Instagram: 1.4 million with a proactive rate of over 96%

On Facebook, we expanded our media matching technology and improved our proactive detection technology in Q2 2021, allowing us to remove elder, violating content. We were able to take action on more infringing content as a result of both of them.

Recent Trends

There were new categories and improvements in reducing prevalence, as well as

  • As we update our proactive detection technology, there were 6.2 million pieces of organised hate content on Facebook in Q2, down from 9.8 million pieces in Q1 2021.
  • Due to a technical fix that allowed us to go back into the past and find content that violated our policies, there are now 16.8 million pieces of suicide and self-injury content, up from 5.1 million pieces in Q1 2021.
  • In the first quarter of 2021, there were 30.1 million pieces of violent and graphic content.

On Instagram in Q2, we took action on:

  • 367,000 of organized hate content, up from 325,000 in Q1 2021
  • 3 million pieces of suicide and self-injury content, up from 2.6 million in Q1 2021
  • 7.6 million pieces of violent and graphic content, compared to 5.5 million pieces in Q1 2021