On 16 October 2026 Apple announced changes to its App Review Guidelines aimed at nonconsensual intimate imagery (NCII) and "bait-and-switch" apps. Two weeks earlier, on 2 October, the Eighth Circuit put Minnesota's anti-nudification statute on hold. The timing is a useful case study: a private gatekeeper acted while a public statute was stuck in constitutional litigation. That is a real advantage of platform rules, though it also carries real risks.
What Apple changed
Guideline 1.2 already told apps that let users create or share content to include filtering, reporting and blocking tools. The existing text lists "nonconsensual intimate imagery" among the challenges and names "AI content creation apps" as in scope. Apple's announcement says moderation requirements now explicitly cover NCII, "including imagery generated using AI tools," as content apps "must proactively prevent and remove." Developers who fail to put safeguards in place, or who don't fix violations after notice, risk removal from the App Store. MacRumors reported that the revised rule could be aimed at apps like xAI's Grok. Apple had warned xAI in April 2026 that it could be removed over nude deepfakes.
The second change is guideline 2.5.19. Apple describes it as clarifying a "longstanding prohibition" on apps that misrepresent what they do during review and then add functionality after approval. Apple's post says apps found doing this face removal from the App Store and from the Apple Developer Program. The post names the target directly: apps that reveal harmful capabilities, "such as 'nudification,'" after approval.
The strongest case for hard law
The best argument for statutes over platform policy is that nudification tools are built to cause a specific harm. A private company's terms can change overnight, apply unevenly and leave victims with no legal right. Minnesota's House File 1606 tried to fill that gap. It passed the House 132-1 and the Senate 65-0 and was signed on 7 May 2026. It took effect on 1 August. It bars operators from letting users "nudify" an image of an identifiable person, defined as generating an intimate part so realistic that a reasonable person would believe it belongs to that person. It also bars advertising such services. Victims can sue for up to three times their actual damages, and the attorney general can seek civil penalties of up to $500,000 per unlawful access, download or use. That is a serious accountability mechanism, and the near-unanimous votes show how little political resistance the goal itself faces.
Why the statute stalled
The Eighth Circuit granted xAI an injunction pending appeal on 2 October. A district judge had earlier found that xAI had not shown irreparable harm. The appellate order did not resolve the First Amendment question, and the underlying case continues. But the episode shows the design problem. A law that regulates a tool, rather than a specific act of harming a specific person, invites a speech challenge over how broadly it reaches. The statute also exempts services that require "the technical skill of a user" to nudify an image, a line that invites argument about where general-purpose image editing ends and nudification begins.
A more durable path already exists. The TAKE IT DOWN Act's removal duties have been enforced by the FTC since 19 May 2026. Covered platforms must remove intimate images and known identical copies within 48 hours of a valid request. The FTC wrote to 15 named platforms, including Apple, Meta, Microsoft and X, to remind them of the obligation. That model attaches duties to a concrete notice about a concrete image. It targets the harm directly and leaves lawful editing and expression alone.
What Apple's approach does well, and where it falls short
Apple's change has three strengths. First, it works at the point of distribution, where a nudification app has to pass review before reaching users. Second, the bait-and-switch rule targets deception, which is a more defensible line than policing a category of software. An app that tells reviewers it is a photo filter and later unlocks a nudifier has misled its gatekeeper, whatever one thinks of the underlying technology. Third, it can change as fast as the abuse does.
The weaknesses matter just as much. Guideline 1.2 asks for "adequate safeguards" and "proactive" prevention, wording that is vague. Developers cannot easily tell what meets the standard, and vague standards push legitimate image-generation tools to over-block. Enforcement is also opaque: there is no public appeals data, no transparency reporting on takedowns and no outside audit. The rules govern one company's store, and web-based services and sideloaded apps sit outside them. As the MacRumors report notes, the changes take effect immediately, but how many existing apps they will remove is not yet known.
A proportionate path
Policymakers should treat platform rules as a complement to law, not a replacement for it. Three principles follow.
- Regulate conduct and notice, not categories of code. Duties triggered by a specific victim's request, as in TAKE IT DOWN, are easier to defend than bans on a class of tool.
- Pair gatekeeper rules with transparency. App stores that remove nudification apps should publish how many they removed, why, and how developers can appeal.
- Draft narrowly. States writing nudification bans should define the harm tightly, such as realistic depictions of identifiable people without consent. They should keep clear exemptions for general-purpose tools and for lawful satire or journalism.
Technology-facilitated gender-based violence falls hardest on women and girls, and a victim does not care whether the fix came from a legislature or a developer program. Apple's move shows that a deception rule and a safeguards duty can deliver protection now. The Minnesota litigation is a reminder that durable protection will depend on laws drawn tightly enough to survive the courts.