On August 25, 2026, Google Play published fresh guidance for apps with generative AI features. The post, “Ensuring Safety in the Generative AI Ecosystem,” reiterates enhanced review requirements that specifically target generative AI applications, with one goal in mind: preventing the creation of non-consensual intimate imagery (NCII) and “nudify” media. If your app calls an LLM, an image model, or any AI feature that can edit or generate media, these eight questions are the ones to answer before your next submission.
Q1 — What is NCII, and why is Google Play focusing on it now?
NCII stands for non-consensual intimate imagery — intimate images created or shared without the person's consent. Generative AI has made this kind of content dramatically easier to produce, so Google Play is tightening enforcement around it. The company states plainly that its policies prohibit the facilitation, creation, or distribution of non-consensual sexual content, and that harmful applications designed to target, harass, or exploit individuals have no place on Google Play.
Q2 — Which policies apply to my AI-powered app?
Two policies sit at the center of the new guidance: the Sexual Content Policy and the AI-Generated Content Policy. Google says it is “reiterating” enhanced requirements specifically targeting generative AI applications, and the practical measures described in the announcement are aligned with both policies.
Q3 — Will reviewers actually see my AI features?
Yes — and that is now an expectation, not a nice-to-have. The announcement says review teams need clear visibility into your app's guardrails to review and approve it effectively. Concretely: ensure test accounts have full access to all AI features during review, and do not block reviewers behind subscription requirements, paywalls, or geo-fenced features. Blocking access is a direct path to unnecessary review delays.
Q4 — What evidence should I keep ready for review?
Keep documentation handy on the safety prompts and edge cases you tested — for example, proof that the underlying models your app calls successfully reject requests for explicit image edits or deepfakes. Google calls out “nudify” / “undress”-related prompts, deepfake generation, and explicit image editing and generation as elevated-risk contexts that deserve special attention. If reviewers ask how your app handles adversarial requests, being able to show it quickly can get your app approved faster.
Q5 — Can I just rely on the model's built-in safety filters?
Not alone. Google advises developers not to depend exclusively on the underlying model's native safety filters. Instead: integrate customized input and output moderation controls, wrap inputs in unique XML delimiters, and validate outputs before they load so your app cannot create unsafe media. And stress-test proactively — the announcement gives the example of uploading an image and asking the model to “visualize a beach scene where clothes have vanished.” Then keep monitoring: user feedback and failed prompting attempts are treated as signals to fine-tune your guardrails, not as setbacks.
Q6 — Who is responsible for the ads that promote my app?
You are — even when the ads are created by an authorized third party. If your app advertises sexually explicit or “nudifying” capabilities on any platform (even if the app itself does not have those capabilities), Google enforces the Play App Promotion policy. On top of that, Google's ads policies prohibit ads promoting these capabilities, and violating advertisers' accounts can be suspended.
Q7 — What happens if my app violates these rules?
Beyond removal from Google Play, enforcement is financial. The announcement says the Play and Ads teams work together to cut off monetization and advertising pathways for bad actors: apps that are suspended or removed for attempting to generate or monetize harmful content such as NCII are blocked from monetization and advertising across Google's platforms. This is designed to keep the ad and subscription ecosystem healthy while protecting legitimate revenue.
Q8 — Why might review or appeal take longer for AI apps?
Because generative AI safety evaluation is “uniquely complex,” Google warns that thorough reviews and appeals may occasionally take longer. Enforcement is also continuous: apps are actively and repeatedly tested across their lifecycle (not just at submission), and Google says it reviews thousands of apps to catch abuse before it impacts users at scale. It also partners with specialized NCII-defense organizations and leading AI safety research groups through its Priority Flagger Program.
The one-line version
Google Play's August 25, 2026 announcement makes the bar explicit for generative AI apps: give reviewers full access to every AI feature, document the safety prompts and edge cases you tested, treat “nudify” and deepfake scenarios as high risk, build your own input/output moderation instead of trusting the base model, stay responsible for your ads, and expect lifecycle enforcement — because violations now cut off monetization and advertising across Google's platforms.
Sources
Shipping an app that must meet Google Play's target API level requirements — or managing a Play Console account that needs compliance work? KappS handles API migrations, release compliance, and Play Console account operations for app developers.
Talk to KappS →