“For the first time we’re introducing a universal design across our platforms.” With those words, Alan Dye, Apple’s VP of human interface design, marked a turning point at WWDC 2025, as Apple unveiled its broadest design update, ever the Liquid Glass UI. The announcement, however, was only the opening act in a keynote that redefined both the look and the intelligence of Apple’s devices.
The Liquid Glass design language, derived from visionOS on Vision Pro, is more than a superficial makeover. It is a technology achievement in dynamic transparency, real-time rendering, and adaptive motion that infuses iOS 26, iPadOS 26, macOS Tahoe 26, watchOS 26, and tvOS 26. Buttons, sliders, tab bars, and sidebars now shine with glass-like translucency, their edges refracting light and reacting to user input as well as ambient conditions. The menu bar in macOS Tahoe 26 is completely transparent now, making the desktop appear more spacious, and iOS 26 menus disintegrate smoothly under touch, recalling the haptic simulations of physical glass.
This design change draws heavily from the conventions of glassmorphism a style characterized by stacked transparency, frosted blurryness, and bright color overlays. Glassmorphism, as visually compelling as it is, comes with real engineering and accessibility challenges. As pointed out in one recent analysis, “it’s almost impossible to reach all the Web Content Accessibility Guidelines (WCAG) criteria and still use Glassmorphism.” Apple’s solution tries to address these issues through adaptive screens and voluntary modes, but aesthetic vs. usability harmony is still a prime point of scrutiny for UI/UX experts.
Aside from design, Apple’s most impactful action might be the release of its underlying AI model Apple Intelligence to third-party developers. Craig Federighi, the senior vice president of software engineering at Apple, clarified, “This work needed more time to reach our high quality bar,” pointing to Apple’s slow, methodical development of its AI features. And now, with the Foundation Models framework, developers can use on-device AI capabilities natively in Swift in just a few lines of code. This allows for experience that is smarter, offline-accessible, and privacy-respecting since all processing can be kept local to the device.
The privacy architecture by Apple is especially impressive. To elaborate, based on detailed technical revelations, Apple Intelligence utilizes a hybrid methodology: lightweight models are executed on-device for everyday tasks, with more demanding requests being sent securely to Private Cloud Compute (PCC) servers. These servers operate with end-to-end encryption, no persistent storage, and public verifiability of deployed code. “Apple notes that it doesn’t train its models with private data or user interactions,” The Verge reported, emphasizing a sharp contrast with industry peers. For tasks requiring external expertise, such as image generation in the Image Playground app, Apple has integrated OpenAI’s ChatGPT, but only with explicit user consent and clear privacy controls.
Apple Intelligence is already apparent throughout the OS. Real-time translation is now integrated into Messages, FaceTime, and even regular phone calls. The system includes English, French, German, Italian, Japanese, Korean, Brazilian Portuguese, Spanish, and Simplified Chinese, and functions even if the other person doesn’t have an iPhone. During a demonstration, after a speaker has spoken, the device provides a conversational translation in the user’s language, with a transcript taking advantage of on-device large language models for speed and privacy.
The new Call Screening feature brings AI onto the front lines of daily interaction. If an unknown caller calls, the iPhone may answer, ask the caller to say their name and reason for calling, transcribe the response, and only ring the user afterwards. This copies Google’s “Hold for Me” but is done with Apple’s privacy-first principle.
Visual Intelligence, Apple’s computer vision system, has also been improved. No longer restricted to images, it now examines anything the iPhone screen is displaying. Screen it with a screenshot, and Visual Intelligence is able to detect objects, pull event information, or even determine similar items for purchase within installed applications. Integration with ChatGPT within Image Playground allows one to create images in certain styles, but Apple makes it a point that “user data would not be shared with OpenAI without a user’s permission,” according to The Independent.
The ramifications for developers are significant. The Foundation Models paradigm facilitates guided generation, tool invocation, and prompt orchestration. Day One apps leverage it for writing assistance with AI, while AllTrails suggests trails through on-device intelligence. The architecture of the system is designed to take advantage of Apple Silicon’s Neural Engine, using state-of-the-art methods like “speculative decoding,” “context pruning,” and “group query attention” for inference efficiency.
The year-named OS system, beginning with iOS 26, also indicates Apple’s intention to consolidate its ecosystem. Cross-platform consistency design and AI feature ensures developers can code once and deploy anywhere, with the confidence that privacy and performance levels are upheld.
For UI/UX designers, the Liquid Glass age calls for new thinking. The interplay of depth, transparency, and motion coupled with the support of real-time rendering provides a more richly textured canvas but also necessitates care about contrast, readability, and accessibility. As glassmorphism shifts from trend to mainstream, the test will be to make sure that beauty is not paid for in the currency of usability.
Apple’s WWDC 2025 announcements have raised the technical and design bar for the industry. The merging of cutting-edge AI, privacy leadership, and a new bold design language constitute a definitive leap in digital experiences on Apple devices.
