A translation, not yet a law
On July 16, the Association for Struggle against Sexual Violence (Cinsel Şiddetle Mücadele Derneği), UN Women, and the Sexual Violence Research Initiative (SVRI) held a webinar to launch the Turkish translation of Global Trends to Prevent and Respond to Technology-Facilitated Violence Against Women and Girls: A Compendium of Emerging Practices — a UN Women/SVRI report cataloguing laws, platform policies, and survivor-support models from around the world (bianet). UN Women policy expert Raphaëlle Rafin told the webinar that digital violence affects roughly half of women worldwide; Hazal Sipahi of the host association put it more bluntly: "the digital realm is not a virtual world separate from the real world, it is real life itself" (bianet). Chayn, the UK-based survivor-support nonprofit, used the same event to preview a Turkish-language version of Bloom, its trauma-informed healing platform.
The timing matters. UN Women's own November 2025 assessment found that 44% of the world's women and girls — 1.8 billion people — live in countries without laws covering digital abuse, and that fewer than 40% of countries criminalize cyberstalking or cyber-harassment specifically (UN Women). The compendium exists precisely because most national penal codes were written before generative AI made fabricated sexual imagery cheap and instant to produce. Turkey is a useful test case for whether translating the playbook into a local language does anything to close that gap on the ground.
What Turkish law actually covers
Turkey has no statute that names AI-generated sexual content, deepfake pornography, or image-based abuse as such. Prosecutors instead route these cases through a patchwork of general-purpose Turkish Penal Code (TCK, Law No. 5237) provisions: sexual harassment (Article 105), threats (106), blackmail (107), persistent stalking (123/A), insult (125), violation of the privacy of private life (134), and unlawful recording or transfer of personal data (135–136) (TCK 5237, mevzuat.gov.tr). None of these was drafted with synthetic imagery in mind. Article 134 punishes invasions of an actual private sphere — a fabricated image that never depicts a real private moment sits awkwardly inside that logic, pushing victims toward blackmail or insult charges that carry lighter penalties and were built for different harms.
On the platform side, Turkey's Internet Law No. 5651 gives victims a notice-and-takedown route: a Criminal Court of Peace can order removal of content that violates personal privacy, and large platforms must respond to such complaints within 48 hours. But 5651 doesn't mention deepfakes by name either — it reaches synthetic sexual imagery only indirectly, through general obscenity and privacy provisions, with no explicit consent standard for AI-generated likenesses (Facia.ai analysis). The result is a system that can act, but inconsistently, and only after the fact.
The case for naming the offense
There's a real argument for closing this gap with a dedicated statute rather than leaving judges to stretch privacy and blackmail law to fit synthetic content. Ambiguity has costs: it makes outcomes unpredictable for victims, gives defense counsel room to argue a fabricated image isn't "private life" at all, and means charging decisions turn on which prosecutor gets the file rather than on a clear legislative standard. The UN Women compendium's push for countries to name technology-facilitated violence explicitly in statute — rather than relying on courts to backfill definitions — reflects a genuine problem, not bureaucratic box-ticking.
Why a broad new offense is the wrong instrument here
But Turkey's specific history argues against solving this with a sweeping new "digital violence" category. Law 5651 has already shown how broadly-worded takedown powers metastasize: bandwidth throttling was used against Twitter, YouTube, and Meta after the 2020 amendments, Wikipedia was blocked for nearly three years, and Discord was blocked in October 2024 under the same architecture (peopleofinternet.com's prior coverage). A Turkish court found in October 2023 that letting the telecoms regulator BTK order removals without judicial sign-off was unlawful — a precedent Ankara should extend, not walk back, if it legislates here.
The proportionate fix is narrower: amend Articles 134 and 136 to explicitly cover non-consensually created or distributed synthetic intimate imagery of an identifiable person, closing the definitional gap without creating a new open-ended content category. Pair it with the existing 48-hour platform response window under 5651, but require judicial — not executive — authorization for any removal order, consistent with the 2023 ruling. That gives victims the clear statutory hook the UN Women compendium is arguing for, without handing BTK a new lever that its record suggests it would eventually point at journalists rather than abusers.
What comes next
A translated compendium and a Turkish-language Bloom platform expand what survivors and advocates can point to when they argue for reform. Neither compels Ankara to act. The gap between the two — global best practice now available in Turkish, and a penal code still running on 2004-era categories — is the actual story here, and it will stay open until parliament, not a webinar, closes it.