AI in Mobile Applications: Google Play Requirements

24.03.2026

Google Play has officially formalised the key requirements applicable to content generated by artificial intelligence (hereinafter, AI). A “simply functioning algorithm” is no longer sufficient. The platform provides that the safety of, and responsibility for, any AI-generated content within an application rests with the developer.

In this article, we will examine precisely what Google Play requires, how these requirements align with global regulation, and what practical steps developers should take in order to ensure that their applications comply with the new standards.

1. The essence of the changes

In 2025, Google Play updated its Developer Program Policy, setting AI-generated content apart as a separate regulated area. The principal requirement is that AI-based applications must not be capable of generating any offensive content. In other words, developers are required to prevent the generation of harmful content in advance, rather than merely responding to complaints. Liability extends to any output produced by the model, including content created by users.

2. Which applications do these rules apply to?

The rules apply to applications in which AI generates content on request, including:

1. Chatbots in which interaction with AI is the primary function.
2. Image generators based on text, visual or voice prompts.
3. Audio/video generators producing deepfakes of real persons.

These rules do not apply to applications with a limited use of AI, including:

• Social networks and platforms where users post AI content but do not generate it within the application itself.
• Applications containing user-generated content, where moderation is governed by policies applicable to user-generated content.
• Applications that briefly describe content not created by AI, such as information from search results or documents (for example, summarising the contents of books), where such concise summarisation is the product’s sole function.
• Productivity tools in which AI helps improve existing functions, for example by creating draft emails in mail applications.

3. What exactly is prohibited?

The core requirement of the policy is the developer’s obligation to prevent the generation of offensive content. Developers must ensure that their generative AI-based applications cannot be used to create materials that violate Google Play’s rules applicable to such content. In particular, the prohibition covers materials involving child exploitation or violence, as well as information capable of misleading users or facilitating dishonest activity.

Google Play identifies the following as specific examples of prohibited materials:

1. Sexual deepfakes created without the consent of the persons depicted;
2. Audio or video recordings of real persons that facilitate fraud;
3. False election-related materials;
4. Content encouraging harm to oneself or others.

4. What must a developer do?

• Implement an in-app complaints mechanism

Applications in which content is created with the help of AI must contain built-in functions enabling users to submit complaints or flag content as unacceptable, so that users can report a violation to the developer without having to leave the application or search for a form on a website.

• Ensure transparency and proper disclosure

Google Play policy does not expressly require disclosure of the use of AI in the application description or documentation. However, the App Promotion policy requires that the description and marketing materials accurately reflect the application’s functionality. If AI is a key feature, failure to mention it may be regarded as misleading.

At the same time, regional legislation (for example, the EU AI Act and the laws of California and New York) expressly requires labelling of AI-generated content and informing users accordingly. Industry practice and user expectations make such disclosure necessary in order to maintain trust and avoid claims.

• Comply with Google Play’s other rules

The policy on AI-generated content supplements, rather than replaces, the basic rules. An application using generative AI must comply with all applicable store policies, for example:

1. Child Endangerment Policy;
2. Deceptive Behavior Policy;
3. Restricted Content Policy.

5. Compliance with international standards

In addition to Google Play rules, regional legislation must also be taken into account:

1. The EU AI Act provides for mandatory labelling of AI-generated content, transparency towards users, and compliance with copyright requirements. For high-risk systems (from August 2026), registration, risk management, documentation and audit will be required.
2. In the United States, regulation is currently based on a “patchwork” approach: federal guidance combined with state laws. However, in December 2025, the President signed the Executive Order Ensuring a National Policy Framework for Artificial Intelligence, thereby launching a unification process. Until a federal standard is established, the laws of key states remain in force. Companies whose applications are available in the United States must therefore comply, for example, with the requirements of New York (disclosure of security protocols and incident reporting), California (content labelling and transparency of training data), and Colorado (combating algorithmic bias).

---

Accordingly, Google Play’s standards on generative AI are creating a new reality for mobile application developers. Google expressly states that it is the developer who is responsible for any content generated by AI within an application and who must prevent the appearance of offensive content in advance.

 

Companies that adapt to the new rules in a timely manner obtain clear advantages: resilience against blocking, reduced legal risks, increased user trust, and the ability to enter international markets without additional barriers. Implementing transparency, content filtering, in-app complaints systems, proper documentation, and compliance with regional legislation is not an expense, but an investment in the product’s long-term stability.

Author: Svetlana Gordey, Daria Gordey.