Facebook saw declines in the amount of content that was detected and removed/flagged as hate speech, according to the platform’s compliance report for December 2021. Content actioned as hate speech dropped to 60,700 pieces from 1,00,000 pieces in November 2021, while its hate speech detection rate slipped to 87.7% from 91.1% in the previous month.
Providing figures for both Facebook and Instagram, the compliance report showed that the former saw a rise in grievances from Indian users while the latter registered a slight dip in grievances compared to last month’s report.
Under the IT Rules 2021, significant social media intermediaries (like Facebook and Instagram) have to publish periodic compliance reports detailing complaints received from users in India and the actions taken by the platform to address these issues. The lack of online content moderation, especially to do with hate speech on Facebook, has been a growing concern and data from compliance reports give an idea on how tech platforms are dealing with such content.
Content actioned by Facebook as spam rises by 2.8 million
Out of the 12 categories of content on which information was provided, the following ones saw the most noteworthy shifts:
Bullying and Harassment: 1,17,000 posts related to bullying and harassment were actioned, up from 1,02,000 posts. The proactive detection rate also increased by 17% (from 40% in November to 57% in December), according to the report.
Spam: Content actioned for spam rose by 2.8 million in December (13.8 million spam posts) over the previous month (11 million spam posts) while the proactive detection rate rose by 0.2%.
Dangerous Organisations and Individuals (Terrorist Propaganda): Content actioned under this category increased to 1,18,300 from 71,700 in November 2021. The proactive detection rate fell to 97.8% from 99.4% in the previous month.
Here, content actioned implies the number of individual pieces of text, photos, videos, and comments which are either removed or labelled with a warning. The proactive detection rate signals the amount of content that was flagged by the platforms’ automated systems without users reporting it.
Small rise in number of grievances sent by Facebook users
User grievances in December rose to 534 from 519 in November and Facebook responded to all of them, the compliance report said. In the case of 436 user grievances, the platform provided resolution ‘tools’ such as channels to report content for specific violations, self-remediation flows whereby users can download their data, avenues to address issues around compromised accounts, etc., and the rest of them went through specialised review. Following the review, 28 grievances had some type of action taken while 67 grievances did not have any action taken.
The result of no action taken happens when a grievance is related to assistance in accessing an account, feedback on a service, etc., the report clarified.
User grievances account for those that are sent to its Grievance Officer or through the contact form set up on its website. Here’s a category-wise breakdown of these grievances for December:
Graphic content actioned increases by nearly 50% on Instagram
Among the ten categories of content under which Instagram provided figures, the following stood out:
Violent and Graphic Content: Content actioned nearly doubled to 6,00,000 pieces from 3,33,400 pieces in November. 99.6% of content actioned in December was proactively detected, up 0.3% from November.
Dangerous Organisations and Individuals (Organised Hate): Content actioned under this rose to 2,500 with a proactive detection rate of 95%. For the last reporting period, the number of content actioned was 1,400 while the proactive detection rate was 87.3%
Child Endangerment (Sexual Exploitation): Content actioned under this category had risen almost 10 times to 12 lakh in the last reporting period. However, the December report shows that this figure has declined to 1,71,200 with a proactive detection rate of 98.3%.
Fewer number of grievances from Instagram users
Instagram received fewer grievances in December 2021 compared to the previous month (424). Out of 414 grievances, the platform provided tools – similar to those provided by Facebook- for 307 grievances and took action in the case of 41 grievances, while the rest saw inaction. Here’s a category-wise breakdown of user grievances to Instagram:
What do the IT Rules 2021 require?
The IT Rules require social media intermediaries to:
- Proactively identify and take down content: This includes content moderation (through automated mechanisms) of posts that are defamatory, obscene, pornographic, paedophilic, invasive of privacy, insulting or harassing on gender, and other types.
- Publish periodic compliance reports: These reports should be published every month and have details of complaints received, action taken, and “other relevant information”.
- Appoint key managerial roles: Significant social media intermediaries (with more than 50 lakh registered users) must appoint a chief compliance officer, nodal contact person, and resident grievance officer, all of whom must be Indian residents and employees of the platform.
- Disable content within 36 hours of government order: The Rules also ask intermediaries to provide information for verification of identity or assist any government agency for crime prevention and investigations no later than 72 hours of receiving a lawful order. They also have to preserve records of disabled content for 180 days.
- Google sees an uptick in complaints from Indian users during December 2021
- Over 1.7 million WhatsApp accounts banned, latest compliance report shows
- Twitter took action against 354 URLs between October and November: Report
Have something to add? Post your comment and gift someone a MediaNama subscription.