Meta Tweaks Policies to Allow Non-Consensual Intimate Imagery Posted to Raise Awareness: Report

Meta has tweaked its adult sexual exploitation policies to allow content that depicts non-consensual sexual touching as long it is to spread awareness, MoneyControl reported yesterday. The content should also not be sensationalist, the victim should be unidentifiable, and no nudity should be involved.

The recent announcement follows a 2022 recommendation from the Oversight Board, the quasi-judicial body overseeing content moderation issues across Meta’s platforms. The Board was reviewing Meta’s decision to restore an Instagram video depicting the sexual assault of a “tribal” woman by a group of Indian men.

“In India, Dalit and Adivasi people, especially women, suffer severe discrimination and crime against them has been rising,” the Board’s verdict from last year noted. “Social media is an important means of documenting such violence and discrimination, and the content in this case appears to have been posted to raise awareness. The post therefore has significant public interest value and enjoys a high degree of protection under international human rights standards.”

Meta is bound by India’s IT Rules to “cause users” not to upload certain kinds of materials—including content that is pornographic, violates bodily privacy, and more. This move by the company indicates how the platform may be balancing its legal obligations globally, against users’ rights to receive and access constructive information on sexual violence.

What prompted the policy change? In 2022, a Dalit rights Instagram account (with 30,000 followers) posted a video depicting men sexually assaulting and harassing an Indian woman, described by the Board and MoneyControl as a tribal woman.

The video was reported by an Instagram user for “sexual solicitation”—human content reviewers went on to find that it violated Meta’s Adult Sexual Exploitation Policy. The policy allows for content that threatens or promotes sexual violence to be removed from Meta platforms. After being flagged by specialists for additional human reviews, the post was restored for its “newsworthiness” (or in the Board’s words, for being in the public interest”). It was also accompanied by an alert warning users of the video’s graphic content (the Board added that users under the age of 18 were restricted from viewing it).

Despite this, Meta went on to refer the issue to the Oversight Board in late 2022. The post raised the question of how to strike an appropriate balance between permitting content condemning sexual exploitation on the platform, and the harms of leaving visuals of sexual harassment up online.

“With regard to its decision to apply the newsworthiness allowance, Meta explained that the public interest value of the content was high because the content was shared by a news organisation that highlights the stories of underrepresented and marginalised populations,” the Board said, while recalling Meta’s submissions in the case. “Meta said the content appears to have been shared with the intent to condemn the behaviour in the video and raise awareness of gender-based violence against tribal women. Adivasi and marginalised voices, Meta argued, have been historically repressed in India and would benefit from greater reach and visibility. Meta also argued that the risk of harm was limited as the depiction did not involve overt nudity or explicit sexual activity, and does not sensationalise. It argued that the “case was exceptional in that the victim’s face is not visible and her identity is not readily identifiable”.

After studying the case, the Board upheld Meta’s decision to restore the video. It also recommended that Meta include exceptions to its policy on adult sexual exploitation—and that depictions of non-consensual sexual touching should be allowed if they intend to raise awareness, do not involve nudity, are not shared to create sensationalism, and the victim is unidentifiable. On August 29th, Meta announced that it had incorporated these changes in its Q2 updates on the Board in May 2023.

Article continues below ⬇, you might also want to read:

• Meta’s Stakeholders Worry About Problematic Content Moderation On Facebook And Other Platforms
• “Security Concerns” Prevented Meta From Releasing Full Report On Impact Of Hate Speech On India
• Meta’s Independent India Human Rights Impact Assessment Summarised In New Report

Still little insight on Meta’s Report on Human Rights Impact Assessment for India: Released last year, the summary of the report was vague about how the platform may be used in India to violate fundamental rights. Critics described the summary as PR fluff, which did little to examine how Meta’s platforms were allegedly fomenting communal and sectarian violence in India.

Similar issues were raised by the Board last year when it delivered its verdict on sexual violence posts last year:

“The Board asked Meta to share its Human Rights Impact Assessment Report for India with the Board, which Meta declined, citing security risks. Meta failed to provide a satisfactory explanation for why sharing the Human Rights Impact Assessment Report with the Board would entail security risks.”

Is the Oversight Board successfully holding Meta to account? Writing in his Substack-newsletter Platformer, Casey Newton recently argued that the Board’s “glacial pace” often leaves harmful content online for much longer than it ‘should’ be. In the case of recent calls for violence by top Cambodian politicians, the Board acted far too slowly, Newton pointed out:

“But when platforms are considering questions related to credible incitement of violence — particularly incitements coming from a head of state — they should resolve them in far less than the 234 days it took the Oversight Board. There are millions of real people who are depending on them. And the board’s members, many of whom were drawn from academia, are still treating the cases referred to them as abstract thought experiments to be debated casually in between graduate seminars.

In Cambodia, the damage is done. The rivals were, in the end, beaten with sticks. Hun Sen’s party “won” the sham election. Last month he said he would soon install his son as prime minister, but would continue to serve as the country’s strongman in chief as the ruler of its sole major political party for years to come.

Earlier this summer, anticipating that he would be banned, Hun Sen briefly decamped to other platforms. He focused his attention on Telegram, with its large user base and famous indifference to content moderation.

But the storm passed, and now the prime minister is posting to Facebook as enthusiastically as ever. Over the past day, his account was updated more than a dozen times — revealing, if nothing else, that at least one player in this platform drama understands the importance of speed and scale.”