Abstract:
Apple recently announced that it has updated its App Store review guidelines and will take stricter management measures against involuntary private images generated by artificial intelligence. Previously, platform rules had banned sexually explicit content, and the latest regulations further require developers to proactively prevent, monitor and remove such deep fake content generated by AI. If developers fail to take effective measures, relevant applications may be removed from the App Store.
In the updated 1.2 clause, Apple clearly names the application categories involving AI content generation for the first time. This move is believed to be related to some generative AI products, including chatbot services that have caused controversy for generating nudity and sexualized deepfake content. Previously, Apple had warned relevant developers that if relevant apps continued to allow the generation of nude deepfake images, they would risk being removed from the App Store. Subsequently, the company added corresponding security protection mechanisms and obtained Apple's approval to release an updated version.
At the same time, Apple has added a new review guideline numbered 2.5.19 to emphasize that developers are not allowed to add new features that were not disclosed during the review stage after the app has passed the review. Apple specifically mentioned so-called "Nudification Apps", which are apps that use artificial intelligence to fake photos of real people into nude or pornographic images. Apple believes that some developers may hide the true purpose of their applications during the review period and wait until the application is approved to go online before enabling related functions.
According to the new regulations, any application containing hidden or undisclosed functions will be banned from the shelves, including applications that deliberately disguise their purpose during the review stage and then open harmful capabilities after approval. Apple pointed out that this type of practice is a typical "bait-and-switch" behavior, and the "undressing" function is a representative case of this type of violation.
In addition to the regulations on AI-generated content, Apple has also revised Part 5 of the review guidelines to extend the original application ban on human trafficking and child exploitation to alternative application distribution channels. This means that apps are subject to the same content security requirements whether they are made available through the official App Store or other permitted distribution methods.
In its announcement to developers, Apple emphasized that the company will not hesitate to take punitive measures such as removing apps from the shelves and terminating developer qualifications for apps and developers that violate these regulations.
This rule update shows that as generative artificial intelligence tools become increasingly popular, Apple is further strengthening its supervision of deep fakes, non-consensual private image generation and related abuses to reduce the risk of AI technology being used to harass, exploit or invade privacy.
Read the full text of the rules:
https://developer.apple.com/app-store/review/guidelines/
Comments