Apple and Google ordered to purge ‘nudify’ apps from App Stores

San Francisco didn't file a complaint. It issued an order. City Attorney David Chiu told Apple and Google to strip dozens of nudify apps — software that digitally undresses people in photos — from their stores within 28 days or face civil penalties. This isn't a takedown request. It's a demand backed by California statutes that criminalize knowingly facilitating non-consensual deepfake pornography and allow victims to sue anyone who helps create it. The platforms knew. They were warned in January. Warned again in April. They kept hosting. They kept processing payments. They kept collecting their 15 to 30 percent cut.

The Tech Transparency Project's April report puts it bluntly: Google and Apple "steered" users toward these apps. Not tolerated. Not missed. Steered. Algorithmic recommendation engines pushed nudify tools to people most likely to pay. That changes the legal geometry entirely. Section 230 immunity was built for passive hosts, not active promoters who monetize the harm. When an app store's recommendation system surfaces a deepfake generator to a teenager who just searched "photo editor," the platform isn't a neutral conduit. It's a participant.

Chiu's office says both companies have been on notice for almost a year about "processing payments for illegal purchases." Let that sink in. A year. TTP's January report identified dozens of apps selling non-consensual intimate images. April's follow-up found them still there, still transacting. Apple and Google didn't just miss the memo. They built the payment rails, approved the updates, indexed the keywords, and took their percentage. Millions in fees, Chiu estimates. The revenue stream creates a motive the law recognizes: reckless aiding and abetting.

The platforms will argue scale. Millions of apps. Automated review. Edge cases. But nudify apps aren't edge cases. They advertise their core function in plain language — "undress anyone," "remove clothes," "deepnude." They charge per image or per month. They exist to manufacture sexual abuse at scale. An automated scanner that flags "calculator" but misses "undress AI" isn't a technical limitation. It's a policy choice. Apple's human review team — the one Cupertino touts as a privacy and safety differentiator — either saw these apps and approved them, or didn't look. Neither defense survives scrutiny.

Google's Play Store has looser gates. That's been its competitive pitch to developers for years. But looseness doesn't excuse illegality. California's 2025 civil statute lets victims pursue third-party facilitators directly. Every woman whose face was stripped onto a porn body without consent now has a legal path to the companies that hosted the tool, processed the payment, and recommended the download. Class actions will follow. Discovery will expose internal emails, review logs, revenue dashboards. The platforms know this.

The deeper problem: nudify apps are the visible tip of a generative iceberg. The same diffusion models that power "artistic" image editors power these weapons. The only difference is the prompt wrapper and the price tag. Apple and Google have positioned themselves as guardians of trusted software ecosystems. They ban malware, spyware, scams, hate speech. They ban apps that violate intellectual property. They ban apps that facilitate gambling in restricted jurisdictions. But they let stand apps whose sole purpose is manufacturing non-consensual sexual imagery — because the revenue was real and the enforcement was optional.

Chiu's letters give 28 days. That's not a negotiation window. It's a statutory predicate. If the apps remain, the city files suit. Civil penalties accumulate per violation per day. The discovery phase alone could cost more than the Apps ever earned. But the real exposure isn't financial. It's precedent. A ruling that app stores bear liability for facilitating deepfake creation rewrites the economics of generative AI distribution. Every platform hosting model weights, every marketplace selling LoRAs, every developer toolchain becomes a potential defendant.

The platforms have one move that matters: purge the category. Not the specific apps named in the letters — the category. Build a classifier that flags any app marketing clothing removal, body modification for sexualization, or deepfake generation of intimate imagery. Ban the keywords. Ban the developer accounts. Refund the subscribers. Publish the takedown logs. Do it before the 28 days expire. Do it without waiting for a court order.

They won't. They'll negotiate. They'll promise improved detection. They'll cite new AI safety hires. They'll remove the exact apps listed and leave the clone army standing. That calculation worked when the harm was diffuse and the law was theoretical. San Francisco just made it concrete. The city doesn't need to prove Apple and Google built the models. It only needs to prove they hosted the storefronts, processed the cards, and recommended the merchandise after being told it was illegal. The evidence already exists in Chiu's files.

Every day the apps stay live, the liability compounds. Every recommendation served, every in-app purchase cleared, every developer payout issued becomes a fresh act of facilitation. The platforms have the technical capacity to end this tonight. They have the legal obligation to end this tonight. They have the moral imperative to end this tonight. What they lack is the will to sacrifice the revenue stream. San Francisco just made that refusal expensive. The only question is whether Apple and Google pay the bill now or pay far more later — in courts, in precedent, in the trust they claim as their core asset.