AI Regulation Tracker  /  Legislation pending

Turkiye bill would make social platforms liable for AI-generated content shared without consent

A draft amendment to Turkiye's Personal Data Protection Law No. 6698 would let regulators fine social-media and digital platforms directly when they permit AI-generated audio, visual, or written content about a person to be shared without consent. It is a pending bill, not yet law.

By Anthony Guerriero  ·  Reviewed by The Leveraged Years Editorial Desk  ·  Published July 7, 2026  ·  Last updated July 7, 2026

A draft law before the Grand National Assembly of Turkiye would shift responsibility for AI-generated content posted without consent onto the platforms that carry it. Submitted in early January 2026, the bill amends Personal Data Protection Law No. 6698, the country's core data-protection statute administered by the Personal Data Protection Authority (KVKK). As reported by several Turkish law firms, it would add a new paragraph to Article 18 of that law, creating administrative liability for social-media services and digital platforms that allow AI-generated audio, visual, or written content depicting a person to be shared without that person's consent.

This is a pending bill, not enacted law. Its text, numbering, and penalty structure could change during parliamentary review, and the reported details below come largely from secondary legal analysis rather than a confirmed final statute.

What the bill reportedly does

The draft targets deepfakes and other synthetic depictions of identifiable people. According to the bill's stated justification, generative AI has increased the risk that a person's image, voice, or likeness can be copied, manipulated, and circulated without consent, harming privacy and reputation. The response is unusual in where it places the burden. Rather than pursuing the individual who uploads the content, the amendment would sanction the platform that permits its distribution. As reported, the fine would reach 5 percent of the platform's turnover from the financial year preceding the year of the shared content, applied separately for each infringing item. Both figures should be read as reported provisions of a draft, not as settled law.

Why platforms, not uploaders

The rationale cited in the bill is that large platforms are not passive conduits. Their ranking and recommendation systems amplify what users post, so the draft treats them as active participants in dissemination. That framing echoes a wider international move away from broad intermediary immunity toward duties that attach to the service itself. For a platform operator, the practical consequence is that a single failure to act on non-consensual synthetic content could, if the bill passes in its reported form, carry a turnover-based penalty rather than a fixed fine.

The placement inside Law No. 6698 also matters. That statute is a data-protection law, and it treats a person's image, voice, and likeness as personal data. By adding platform liability through Article 18, which already houses the law's administrative sanctions, the drafters route the new duty through an existing enforcement structure administered by the KVKK rather than building a standalone regime. For counsel, that means the familiar concepts of consent, data subject, and administrative fine would carry over, even though the trigger is new.

What it does not do

The amendment, as described, does not ban AI-generated content, and it is not a general AI statute. It is a targeted addition to a data-protection law focused on non-consensual depictions of individuals. It is also distinct from Turkiye's separate, broader draft AI Law (Bill No. 2/2234), which addresses artificial intelligence governance more generally. Readers tracking Turkiye should treat the two as parallel but separate legislative tracks. Because the bill is not yet law, no enforcement, effective date, or implementing regulation exists today.

The cross-border angle for US readers

A US firm is not bound by a Turkish bill that has not passed. The relevance is operational. Major social platforms subject to this measure are often US-headquartered, and a turnover-based fine assessed per item creates exposure that scales with reach, not with the number of violations a firm considers material. Trust-and-safety and legal teams that operate globally tend to build to the strictest applicable standard, so a Turkish platform-liability rule, if enacted, could influence how consent and takedown are handled well beyond Turkiye. Marketers using synthetic media in Turkish campaigns would also need documented consent from any depicted person.

What to watch

The load-bearing questions are whether the bill advances, whether the reported 5 percent per-item penalty survives committee, how "consent" and "platform" are defined in the final text, and what notice mechanism the law requires before liability attaches. The definition of a covered platform will decide how far the rule reaches, and the point at which liability attaches, whether on posting or only after a valid notice, will determine how heavily it falls on trust-and-safety operations. Until the enacted text is published, the responsible posture is to plan against the reported framework while treating its specifics as provisional, and to verify each figure against the final statute rather than against law-firm summaries of a draft.

Browse the full AI Regulation News tracker

Informational analysis for working professionals, not legal advice. Confirm how any rule applies to your situation with qualified counsel.