Content warning: This piece discusses content moderation, non-sexual nudity, and references to sexually explicit advertising.
Meta says it enforces one set of rules for everyone. But when non-sexual images of Black Brazilian life get flagged as ‘adult’, platform nudity bans aren’t just moderating content – they’re governing visibility.
In March 2023, the official Instagram account for a Brazilian political party posted an image of Yanomami women in traditional clothing, praising efforts to stop illegal mining on Yanomami lands. The post did what political comms is meant to do: translate policy into a moral picture, then push it through the feed before attention moves on. It was reported, routed through Meta’s moderation system, and removed under the company’s Adult Nudity and Sexual Activity rules. The account later got the post restored only after the administrator appealed through a personal contact at the company, and Meta re-labelled it as ‘newsworthy’ (Meta’s Oversight Board).
That sequence matters for Afro-Brazilian organisers, artists, educators and cultural workers because it shows how ‘nudity’ travels as a shortcut label. It can slide from computer vision to a strike system to a restriction, and then to a closed-door exception, all while the public story remains: we enforce the same rules for everyone. It’s the same logic of automated “neutrality” that shows up in border tech—see Facial recognition at borders and the racialised traveller—where supposedly universal rules still produce uneven, racialised outcomes.
In June 2025, Meta’s Oversight Board reviewed a bundle of four ‘adult nudity’ cases and overturned Meta’s decisions in three of them. One of those four cases involved Instagram in Brazil; the rest spanned other locations and, in one instance, Facebook. Taken together, the bundle shows how ‘nudity’ enforcement can end up governing political and cultural visibility – especially when non-sexual images of racialised bodies are treated as inherently adult (Meta’s Oversight Board).
Platform nudity bans: the notice that stops a meeting
The first violence of platform nudity bans is their timing. They land mid-task, between drafting a caption and replying to a volunteer, as a door slammed in the face of routine. The notification is brief, almost boring, and it rarely matches the social context of the image. You can feel your organising calendar shrink in real time. You’re left negotiating with a system that offers rules without conversation — and treats interruption as normal.
For Afro-Brazilian activism, ‘skin’ is not a neutral surface. It’s a contested archive shaped by slavery, religious persecution and the long policing of Black joy in public space. When platform nudity bans treat a Black body as inherently adult, they repeat a centuries-old suspicion in new paperwork. The platform may call it safety. Communities experience it as a familiar kind of moral surveillance.
The Oversight Board’s case bundle shows how removals can happen even when content is explicitly non-sexual. In the Brazil case, the image was political communication, not erotica, and it still triggered a nudity-policy pathway. Meta ultimately let it stay only through a ‘newsworthiness’ allowance applied after the fact. The decision serves as a reminder that exceptions are an inherent part of governance, rather than a fix for mistakes.
A strike system turns that governance into discipline. Once a strike is logged, your future posts travel under a shadow you can’t fully see. You post with a new kind of caution: cropping, blurring, or avoiding certain angles — the quiet workarounds platform nudity bans train into routine. Over time, the audience learns the code too, reading your euphemisms like a second language. That’s how censorship becomes a shared habit.
The labour of appealing is also political. It’s time that could have been spent on mutual aid, rehearsal, childcare coordination or community defence. It’s often done by the same people who already carry invisible admin work in collectives. When appeals fail without explanation, the platform isn’t simply rejecting content. It’s declining to recognise the organiser as a rights-bearing user.
In practice, takedowns don’t only affect speech. They affect income for cultural workers who sell tickets, teach classes or promote gigs through Instagram. They affect safety channels for communities that rely on private messages when public space is risky. They affect cultural memory because posts function as informal archives. When content disappears, so does the breadcrumb trail of a movement.
We should be precise about what’s being argued here. The claim is not that every nudity rule is illegitimate. The claim is that enforcement patterns can become racialised through design choices, classifier errors, and unequal access to exceptions. That’s why ‘content moderation double standards’ is the right phrase, not ‘the algorithm went rogue’. Accountability sits with the platform’s system and its policy decisions (Meta’s Oversight Board).
The urgency isn’t theoretical in late 2025. The Guardian reported a global wave of removals and severe restrictions hitting reproductive health and queer organisations across Meta platforms, with campaigners describing opaque reasoning and inconsistent reinstatements. That story is about health content, not Afro-Brazilian protest – but it points to the same structural problem: rules applied at scale, with weak explanation to the people most affected. If this can happen to well-known NGOs, grassroots cultural organisers are even more exposed (The Guardian).
What the platform calls ‘nudity’
‘Instagram nudity takedowns’ sounds like an obvious category until you ask what counts. Platforms often define nudity through body parts, not meanings, because body parts are easier to detect at scale. That makes sense for a classifier, but it flattens culture into anatomy. It also inherits gendered and racialised assumptions about whose body is ‘sexual by default’. The rules can become a moral border drawn across the feed.
The Oversight Board’s June 2025 decision spells out this flattening. Posts showing Indigenous women with visible nipples in culturally specific contexts were treated as policy violations, even when there was no sexual framing. In two Namibia cases, automation removed content and human review confirmed the removal, reinforcing the initial misclassification. In the Brazil case, the post was removed and later restored via ‘newsworthiness’. This is a pipeline problem, not a one-off error (Meta’s Oversight Board).
Afro-Brazilian organisers recognise the pattern because Afro-Brazilian visual culture often involves the body in public space: carnival, blocos afro, dance, percussion, and religious ceremonies where clothing and movement carry ancestral meaning. These aren’t ‘adult themes’ in community terms, even when skin is visible. Yet the platform’s standard can treat them as inherently suspicious. Visibility becomes conditional on conformity to a narrow aesthetic.
When people say ‘the classifier is biased’, they often skip the mechanism. Classifiers are trained on labelled data, and labels reflect what moderators and institutions have historically marked as sexual or indecent. If training data over-represents some bodies as adult content, the system learns that association. That’s what ‘racialised enforcement’ can look like without any single moderator deciding to be racist. Power hides in aggregation.
User reporting adds another layer of vulnerability. Brigading doesn’t need to be highly organised to be effective. It can be a loose swarm of moral outrage, often aimed at women, queer people and Black creators. Even a small reporting wave can push content into review queues. The result is that harassment tactics can be laundered into ‘policy enforcement’.
Strikes and restrictions are where the harm becomes durable. A removed post is a momentary loss, but a restricted account changes future reach. The platform rarely provides an intelligible strike history in a way that helps users plan. It becomes difficult to tell whether you’re being punished, downranked, or ignored. That uncertainty is a governance tool because it encourages self-censorship.
Appeals are supposed to be the human backstop, but the Oversight Board example shows how uneven that backstop can be. In the Brazil case, the user appealed through a personal contact and got a different outcome than standard routes tend to deliver. That’s not a story about individual luck. It’s a story about unequal access to remedy – a fairness issue in itself (Meta’s Oversight Board).
If you want a practical definition of appeals transparency, it’s this: can a user see what rule was triggered, what evidence the platform relied on, and what to do differently next time? Without that, appeals are theatre. The user is asked to perform compliance without being told the criteria. In communities already stereotyped as indecent, that silence is not neutral.
How the ‘nudity’ label travels
From pixels to penalties
The pipeline starts with detection. In plain terms, a system scans images and video frames for shapes and patterns associated with nudity. It produces a confidence score, not a moral judgement. That score can trigger automatic removal, or it can send content to human review. Either way, the initial classification shapes what happens next.
Meta’s own integrity reporting suggests how measurement and enforcement can shift over time. In its Integrity Reports for Q2 2025, Meta noted that the reported prevalence of adult nudity and sexual activity appeared to increase on Instagram due to improvements in measurement. That kind of statement is easy to read as a technical footnote. It’s also a reminder that what the company ‘sees’ depends on its tools, not only on user behaviour. Measurement choices affect enforcement priorities (Meta Transparency).
Detection isn’t isolated from incentives. Platforms want to reduce certain harms, but they also want to avoid scandal and regulatory penalties. Nudity policies are reputationally sensitive because they touch children, advertisers, and moral panic. That makes ‘nudity’ a category platforms often enforce aggressively. The risk is that aggressive enforcement becomes blunt enforcement.
User reporting is the social layer that can amplify classifier mistakes. A report is a form of attention, and attention is a scarce resource inside moderation operations. When a post is reported, it can be prioritised regardless of its cultural context. That means a conservative viewer’s discomfort can shape what content gets reviewed first. The platform’s system turns social power into moderation input.
Human review doesn’t automatically solve the context problem. Moderators work under time pressure and must apply rules consistently across cultures. They often have limited local knowledge of Afro-Brazilian religions, protest aesthetics or the politics of the body. A reviewer may see bare skin and default to prohibition. Context is expensive, and speed is rewarded.
The penalty ladder turns classification into consequence. A single removal may not matter, but repeated removals can trigger restrictions. Restrictions can be temporary, but they often hit during crucial campaign windows. The organiser experiences this as targeted disruption even when the platform insists it is standard enforcement. Scale makes personal impact easy to dismiss.
‘Newsworthiness’ and ‘public interest’ exceptions can be a safety valve – but they can also be a power valve. In the Oversight Board bundle, Meta restored the Brazilian post under ‘newsworthiness’ after initially removing it. That shows the exception exists, but it also shows it’s not reliably accessible. When exceptions are ad hoc, the platform becomes an opaque editor (Meta’s Oversight Board).
This is why the debate shouldn’t stop at ‘AI fairness’. Fairness metrics aren’t the same as accountability. A system can be statistically improved and still reproduce cultural hierarchies if the underlying policy treats certain bodies as more indecent. For Afro-Brazilian communities, the question isn’t only whether the model performs well. It’s whether the platform understands the body as culture, not just content.
Double standards without false equivalence
The clearest way to show double standards is not to argue that all nudity should be allowed. It’s to show inconsistency inside the platform’s own stated goals. If ‘adult content’ is the concern, then explicit sexual material should be reliably removed. Yet investigations keep finding major enforcement gaps in paid content ecosystems. That contrast makes grassroots removals harder to justify.
In January 2025, Le Monde reported on an AI Forensics investigation that found thousands of pornographic ads circulating on Facebook and Instagram despite Meta’s policies. The reporting described how explicit ads could evade review and reach large audiences, while similar imagery posted by ordinary users was removed quickly. That pattern suggests stricter enforcement where it’s cheap – and weaker enforcement where money flows (Le Monde).
The pattern matters for Afro-Brazilian cultural workers because many rely on unpaid distribution. A community dance teacher posting a performance clip is not buying ad inventory. A religious educator posting a ceremony image is not paying for reach. If enforcement is harsher on organic posts than on ads, the burden falls on those with less capital. The platform becomes a market where visibility is safer when it is monetised.
Double standards also show up in access to escalation. In the Oversight Board’s Brazil case, the user appealed via a personal contact and secured a reversal with a label. That implies a tier of influence sitting above normal appeals. It’s not only unfair; it’s corrosive. Communities learn that rights depend on networks, not rules (Meta’s Oversight Board).
Regulatory pressure is meant to push platforms towards clearer reasoning. In November 2024, the European Commission adopted an Implementing Regulation to harmonise transparency reporting under the Digital Services Act (DSA). Templates won’t solve cultural bias on their own, but they can make patterns harder to hide – and easier to compare across platforms (European Commission, Digital Strategy).
Civil society groups keep warning that transparency without detail becomes spin. In February 2025, the Center for Countering Digital Hate (CCDH) published More Transparency and Less Spin, arguing that Meta’s enforcement communications can be unclear about what receives proactive enforcement versus what is left to user reporting. That distinction matters because reporting systems can be gamed through harassment. Where enforcement relies on reports, brigading dynamics gain leverage (CCDH).
Brazilian digital rights debates add another piece: local harm, global rules. In April 2025, Aláfia Lab published Do Meme ao Ódio: Como o Racismo se Manifesta nas Plataformas Digitais, emphasising the limits of content moderation in addressing racism. When racist content stays up, and cultural content is removed, communities experience a perverse prioritisation (Aláfia Lab).
This is where comparison without false equivalence matters. Afro-Brazilian organisers aren’t asking for pornographic material to be tolerated. They’re asking why cultural, educational and political images get treated as adult while genuinely exploitative material can slip through. They’re asking why the platform can be nuanced for advertisers but crude for communities. That’s a governance question, not a morality debate.
The politics of the body
Afro-Brazilian skin as heritage, not content
Afro-Brazilian body politics can’t be separated from history. The Black body in Brazil has been framed through forced labour, racial ‘science’, and a long afterlife of sexual stereotypes. Those stereotypes don’t vanish when a post is uploaded. They can be reactivated by a policy that treats certain anatomy as inherently adult. A classifier isn’t born in a vacuum.
Afro-Brazilian cultural expression also uses the body as language. In blocos afro, dance and percussion carry stories of resistance, mourning and celebration. In Candomblé and related traditions, clothing, gesture and adornment can mark spiritual roles and communal belonging. Skin may be visible because the culture doesn’t read it as indecent. A platform that only reads skin as risk is missing the grammar.
The Oversight Board’s decision about Indigenous women isn’t an Afro-Brazilian case, but it is a Brazil-relevant mirror. It shows how non-sexual nudity tied to belief and custom can be misread by global rules, and how exceptions are granted unevenly. Afro-Brazilian communities live with similar misreadings, especially when imagery sits outside Western modesty norms. The point isn’t to collapse Indigenous and Afro-Brazilian experiences into one story. The point is to show how racialised bodies can be routed through suspicion (Meta’s Oversight Board).
Protest aesthetics often deliberately challenge respectability. Black feminist movements have long questioned why dignity is defined through modesty, especially when modesty has been demanded by those who sexualised Black women in the first place. When a protest image uses the body to reject shame, the platform can treat it as ‘adult content’. That’s not a neutral misclassification; it’s a political disagreement disguised as policy enforcement.
There’s also a class dimension. People with access to galleries, press coverage and institutional legitimacy can get their images framed as ‘art’ or ‘news’. People organising in neighbourhood collectives often can’t. Their posts are seen without protective context and without media amplification. The platform’s exceptions can become a privilege that tracks existing inequality.
The risk of flattening context grows when platforms optimise for speed. Moderation at scale rewards quick decisions and clear categories. Afro-Brazilian cultural expression is rarely ‘clear’ in that sense, because it carries layered meanings across Africa, Brazil and the diaspora. A rulebook that can’t hold layered meaning will push communities towards simplification. That simplification is a cultural loss, not only a speech loss.
When content is removed, it can also feed external stigma. A takedown can be screenshotted and circulated as ‘proof’ that a community is indecent. In conservative climates, this can escalate harassment or justify further policing. The platform’s decision becomes a weapon in offline conflict. That’s why takedowns have safety implications beyond the feed.
To name this ‘racial bias’ isn’t to claim intent without evidence. It’s to describe an outcome pattern shaped by stereotypes, operational incentives and uneven access to remedy. It’s also to insist that platforms are cultural actors, not neutral pipes. Their rules create boundaries around what kinds of bodies can be safely visible. In Brazil, those boundaries intersect with race in predictable ways.
The cost of disappearing archives
Material consequences are often dismissed as ‘just social media’, but organisers know better. A restricted account can mean a cancelled workshop, because people simply don’t see the announcement. It can mean less money for transport stipends, because the fundraiser reaches fewer donors. It can mean silence in a crisis, because direct messages slow down or accounts are frozen. Visibility is infrastructure now.
The Guardian’s reporting on Meta restrictions against abortion and queer organisations shows how quickly infrastructure can be disrupted. The article described dozens of account removals and severe restrictions beginning in October 2025, with users receiving vague reasons and inconsistent reinstatements. It also highlighted how campaigners track incidents across regions, suggesting a broader pattern rather than isolated errors (The Guardian).
Cultural memory is another casualty. Instagram posts function as living archives of performances, protests and community teaching. When posts are removed, the archive becomes patchy and unreliable. Younger activists lose access to visual lineage, and historians lose material evidence. A movement that can’t keep its own receipts is easier to erase.
There’s also the psychological cost of constant negotiation. Creators learn to anticipate punishment and adjust their work accordingly. That can blunt the sharpest parts of protest art, which often relies on bodily presence. It can also reshape how communities see themselves, internalising the platform’s suspicion. Self-censorship isn’t only silence; it’s a changed self.
The economic cost isn’t evenly distributed. Influencers with diversified income streams can survive a restriction. Grassroots educators, community photographers and cultural organisers often can’t. A single month of reduced reach can mean rent stress or project collapse. Platform risk becomes livelihood risk.
Transparency reporting is often presented as proof that companies are improving. Yet headline numbers don’t reveal whose content is removed, whose appeals succeed, or whose cultural context is routinely misunderstood. A report can be technically transparent and still leave communities in the dark about the logic that governs them (Meta Transparency).
When pornographic ads slip through, the injustice becomes harder to ignore. Le Monde’s account of thousands of explicit ads circulating unmoderated points to a gap between policy claims and enforcement practice. If explicit ads can circulate while protest images are flagged, communities are right to ask whose safety is being protected (Le Monde).
All of this adds up to a simple truth: moderation decisions redistribute power. They decide which communities can mobilise quickly and which communities must always fight the platform first. They decide whose bodies are legible as culture and whose bodies are reduced to adult content. They decide whose archives can accumulate and whose are routinely punctured. That’s why Instagram ‘nudity’ takedowns are political events.
Accountability beyond PR
What meaningful change would look like
Accountability starts with reasons, not vibes. Users need granular explanations that connect the specific content to the specific rule and the specific enforcement step. They need strike histories that can be reviewed and contested. They need appeal outcomes that include logic, not only verdicts. Without that, appeals are a closed ritual.
Regulators are pushing platforms towards more standardised transparency, at least in Europe. The European Commission’s November 2024 move to harmonise transparency reporting templates under the Digital Services Act is one attempt to force comparability. Templates can’t solve cultural bias, but they can make patterns harder to hide – and support external scrutiny by researchers and civil society (European Commission, Digital Strategy).
Independent review must be accessible, not only symbolic. The Oversight Board exists, but its scale is limited compared to everyday harms. Even within that system, the Brazil case shows how exceptions and escalation can be uneven. The Board recommended making a clear, public exception for culturally contextual, non-sexual Indigenous nudity – rather than relying on ad hoc permissions. The principle applies more broadly: publish criteria, and apply them consistently (Meta’s Oversight Board).
Platforms also need to be honest about what they enforce proactively. CCDH’s More Transparency and Less Spin argues that unclear communication about proactive detection versus reliance on user reports matters because reporting systems can be gamed. A platform that leans on reports for some harms and automation for others is making a political choice about what it treats as urgent (CCDH).
Brazilian civil society is also calling for deeper participation by those most harmed. Aláfia Lab’s reporting underlines how moderation systems can fail to curb racist content while communities face other forms of policing. If a platform can’t reliably address racism in Portuguese-language contexts but can aggressively detect skin, the outcome is a skewed safety regime. Meaningful change would prioritise anti-racist enforcement alongside careful cultural context handling. It wouldn’t treat ‘nudity’ as the easy win (Aláfia Lab).
The goal isn’t to demand that platforms become perfect cultural anthropologists. The goal is to demand that they stop pretending context is optional. Afro-Brazilian cultural and protest imagery should have clear policy pathways for educational, documentary, artistic and political value. Those pathways should be available through ordinary appeals, not personal contacts. When the system makes mistakes, the remedy should be a right, not a favour.
A collective field kit for organisers
Document everything as if you’ll need it later, because you probably will. Save the takedown notice, the reason code, the timestamp, and the exact content that was removed. Save your caption text and any relevant context that explains why the image is political or cultural. Keep a simple log of strikes and restrictions in a shared collective document. Treat this as movement record-keeping, not admin busywork.
Build parallel archives outside the platform. Keep local backups of images and videos in shared drives with clear folder structures. Publish key material on at least one website you control, even if Instagram is the distribution channel. Use web archiving tools when safe and appropriate, and save screenshots with dates. If the platform removes a post, you should still be able to prove it existed.
Design captions to carry context without apologising for your body. Name the cultural or political setting plainly, because classifiers can’t read meaning from silence. Avoid code words that might look like adult solicitation to automated systems. Use educational framing when it’s true, not as a disguise. Context should be a description of reality, not a performance of respectability.
Plan for reporting waves as an operational risk. Watch for sudden spikes in hostile comments or mass-reporting threats. Rotate admin access, secure accounts with strong authentication, and define who handles appeals before a crisis. Keep contact lists for supportive journalists, lawyers and digital rights organisations. Solidarity is faster when it is pre-organised.
When you appeal, be specific and brief. Identify the relevant policy exception in plain language, such as documentary, educational, political or cultural context. State clearly that the content is non-sexual, and explain what community practice it represents. Attach evidence where possible, such as links to reputable coverage of the event or tradition. Save the appeal text and the response for future pattern-building.
Diversify platforms, but don’t romanticise ‘just leaving’. Instagram may be where your community already is, and abandoning it can mean abandoning people. Diversification should be about resilience, not purity. Use newsletters, messaging lists, community radio and in-person networks alongside social media. The aim is to reduce single-point failure, not to shame people for using the dominant tool.
Finally, treat ‘visibility’ as a collective right, not an individual brand problem. When one Afro-Brazilian organiser is hit with an Instagram ‘nudity’ takedown, it’s rarely only about one post. It’s about how the platform imagines who is allowed to be seen without being sexualised. The work ahead is cultural, technical and political at once. It will require receipts, solidarity, and a refusal to accept suspicion as the price of presence. You may be interested in this article.
Sources
Meta Oversight Board, ‘Images of Partially Nude Indigenous Women’ (3 June 2025).
Meta Transparency, ‘Integrity Reports, Second Quarter 2025’ (27 August 2025).
European Commission, ‘Commission harmonises transparency reporting rules under the Digital Services Act’ (4 November 2024).
Center for Countering Digital Hate (CCDH), More Transparency and Less Spin (24 February 2025).
Le Monde, ‘Thousands of pornographic ads go unmoderated on Facebook and Instagram’ (8 January 2025).
The Guardian, ‘Meta shuts down global accounts linked to abortion advice and queer content’ (11 December 2025).
Aláfia Lab, Do Meme ao Ódio: Como o Racismo se Manifesta nas Plataformas Digitais (April 2025).
Keep Independent Voices Alive!
Rock & Art – Cultural Outreach is more than a magazine; it’s a movement—a platform for intersectional culture and slow journalism, created by volunteers with passion and purpose.
But we need your help to continue sharing these untold stories. Your support keeps our indie media outlet alive and thriving.
Donate today and join us in shaping a more inclusive, thoughtful world of storytelling. Every contribution matters.”
