On May 14, 2026, the Turkish outlet bianet reported on a webinar launching the Turkey special edition of Global Information Society Watch (GisWatch), "Deepening Digital Oppression in Turkey and Feminist Resilience." Co-produced by the Association for Progressive Communications (APC), the Combat Sexual Violence Association (CŞMD), and Genderscope, the report documents technology-facilitated gender-based violence (TFGBV) against women and LGBTQ+ activists — coordinated doxxing, death threats, and sexualised harassment that spikes after flashpoints like International Women's Day and Pride marches — and argues that algorithmic moderation on platforms such as X and Google compounds the harm by quietly throttling feminist and independent media.
The documentation of abuse is credible and important. Where the report's framing needs sharpening — from a free-expression standpoint — is its tendency to treat opaque platform algorithms and overt state censorship as roughly equivalent threats. They are not. In Turkey, the dominant mechanism silencing feminist voices is not a recommendation engine. It is the state.
The abuse is real and measurable
The GisWatch authors are not describing a fringe problem. A December 2025 report led by TheNerve with City St George's, the International Center for Journalists, and UNESCO, released by UN Women, found that 70% of surveyed women human rights defenders, activists, and journalists had experienced online violence in their work, and 41% reported offline harm linked to that online abuse. Nearly a quarter reported AI-assisted attacks such as deepfake imagery — a category that barely registered five years ago.
This is the strongest case for platform responsibility, and it deserves to be stated plainly: when doxxing exposes an activist's home address or coordinated accounts issue credible death threats, that is not "speech" to be weighed against "more speech." It is targeted intimidation designed to drive women and LGBTQ+ people out of public life. Platforms that profit from hosting these networks have a legitimate, proportionate duty to detect, de-amplify, and remove them swiftly — and to give targets fast, human-reviewed recourse. A pro-innovation position that ignored this would be hollow.
But the heavier hand is statutory, not algorithmic
The difficulty with blaming "the algorithm" for silencing feminist media is that Turkey already has a far blunter and better-documented instrument: Law No. 5651, the 2007 internet law, and especially its Article 8/A emergency-blocking power.
In a May 8, 2025 open letter, ARTICLE 19, the Freedom of Expression Association (İFÖD), and Human Rights Watch documented that Turkey's Information Technologies Authority (BTK) used Article 8/A to block 471 X accounts between March 19 and April 12, 2025, affecting roughly 17.2 million followers. Among them were at least 22 women's rights groups with around half a million followers, alongside student organisations and journalists — accounts the signatories said were targeted to "choke protest coordination and feminist dissent." This is censorship with a case number, not an inscrutable ranking change.
The legal architecture compounds the problem. As the Council of Europe's Venice Commission warned in its 2016 opinion on Law 5651, the statute grants administrative bodies sweeping discretion to block access with minimal judicial scrutiny. Turkey's own Constitutional Court, in a decision published in the Official Gazette on January 10, 2024, annulled parts of the law's content-removal provisions as unconstitutional — yet left the administrative blocking power substantially intact. The result is a system in which a regulator can vanish a feminist collective's reach overnight, and the affected account has almost no practical way to contest it.
This matters for how we read the GisWatch finding. When feminist media loses visibility in Turkey, the first-order cause is usually a lawful-on-paper blocking order, court-ordered geographic restriction, or the chilling effect of prosecution risk — not a Silicon Valley A/B test. Conflating the two risks letting the actual censor off the hook.
The proportionate fix: transparency and due process, not new state leverage
The report's instinct — that platform opacity is a problem — is sound. Activists genuinely cannot tell whether a sudden drop in reach is organic, an enforcement action, or compliance with a secret government request. The answer, though, is not to hand the Turkish state more authority to define and police "harmful" content; that power would predictably be turned against the very feminists the report defends. Turkey's history of stretching "public order" and "national security" to cover dissent makes that risk concrete.
The proportionate path runs the other way:
- Transparency by default. Platforms should notify users when a government order — not a policy violation — restricts their content, publish per-country removal and blocking data, and label state-compelled actions. The EU's Digital Services Act model of statement-of-reasons and audited transparency reporting is a workable template, achieved without empowering an authoritarian regulator.
- Faster, human recourse for TFGBV targets. Anti-abuse duties should be measured by how quickly genuine doxxing and threats are actioned and how reliably wrongly-suppressed activists are restored — not by raw takedown volume, which rewards over-removal.
- Resisting blocking-order overreach. X's own decision to comply with but legally challenge a 2025 court order restricting Istanbul Mayor Ekrem İmamoğlu's account shows platforms can push back. That posture — comply minimally, litigate, disclose — should be the floor, not the exception.
GisWatch 2026 performs a real service in naming the gendered texture of Turkey's digital repression. The most useful response is to keep the two threats analytically distinct: hold platforms to account for the abuse they amplify and the suppression they obscure, while keeping the brightest spotlight on the statute that lets the state switch off feminist speech with a single order.