Apple has introduced several new AI image generation features in its latest iOS update, but with some key guardrails to prevent misuse and misrepresentation.
AI-generated images have a troubled track record, with deepfakes being a major concern. However, these issues go beyond just deepfakes. There were recent incidents involving Google’s Pixel Studio app, which created AI-generated images that were insensitive and disturbing.
To address these concerns, Apple has implemented several safeguards in iOS 18.2. Firstly, the company is limiting the realism of its image generation features to prevent photorealistic images from being created. Genmoji and Image Playground can only produce animated or illustrative-style images in one of three styles: Animation, Illustration, and Sketch.
Secondly, Apple is blocking recognized IP characters from being used in AI-generated images. This means that attempting to create an image of a character like Darth Vader with a Santa hat on will result in an error message.
Thirdly, Apple has introduced a waitlist feature for its new AI image tools. This is a progressive rollout, meaning that even if you have access to these features in iOS 18.1, you’ll need to explicitly request access to them in iOS 18.2 by opening the Image Playground or Genmoji app.
These guardrails are part of a larger trend of Apple taking a more cautious approach to its AI features. With several new AI capabilities set to be released next month, including ChatGPT integration and Visual Intelligence on iPhone 16 models, it’s clear that Apple is prioritizing responsibility over innovation in its pursuit of artificial intelligence.
Source: https://9to5mac.com/2024/11/14/ios-182s-genmoji-and-image-playground-have-three-key-guardrails-you-should-know