Apple rumored to drop systemwide live translation at WWDC 2025
Apple is set to expand Apple Intelligence at WWDC 2025 with major upgrades to translation and developer tooling, according to Bloomberg. Users could get automated Message translation, live phone call translation, and real-time conversation translation right through AirPods, alongside new APIs opening Genmoji-style tools to third-party apps.
If you spent the last year waiting for the all-powerful, deeply personalized Siri promised at WWDC 2024, you will have to keep waiting.
At WWDC 2025 on Monday, Apple unveiled its latest wave of Apple Intelligence updates alongside iOS 26 and its next-generation operating systems. Instead of doubling down on assistant hype, the company shifted focus toward pragmatic upgrades: system-wide live translation, on-screen visual context, and direct developer access to on-device models.
Translation across calls and eyes on your screen
The biggest day-to-day upgrade for everyday users centers on translation.
Inside Messages, Apple’s on-device AI will automatically translate outgoing text into your recipient’s native language in real time, and translate replies right back as they arrive. During FaceTime chats, the system displays live subtitles, while regular phone calls now feature voiced live translation. Reports ahead of the keynote also pointed to AirPods-enabled live conversation translation with automatic language detection.
Visual Intelligence is also breaking out of the camera viewfinder. By pressing the standard screenshot button combination on an iPhone, users can now scan whatever is displayed on their screen.
Spot an event flyer on social media? iOS 26 suggests adding it directly to your calendar. You can reverse-search products across Google or Etsy, or feed the screen capture into ChatGPT to ask questions about what you are looking at. Like previous implementations, data is only routed to OpenAI if you grant permission.
On the playful side, Genmoji now lets you fuse two Unicode emojis into an entirely new sticker, like mashing a sloth with a lightbulb. Image Playground adds support for custom group chat backgrounds and new visual styles through ChatGPT.
Three lines of code for developers
For developers, Apple finally cracked open its on-device foundation models through the new Foundation Models framework.
Apple claims developers can tap into local intelligence with as few as three lines of Swift code. The inference runs entirely on-device, completely offline, and without any API fees. Apple highlighted Kahoot! as an early example, showing how the learning app could generate personalized quizzes locally.
Power users also get an overhauled Shortcuts app. Automated workflows can now plug into on-device models, Private Cloud Compute, or ChatGPT to generate responses, such as cross-referencing lecture audio transcripts against student notes.
Even Apple Wallet received an AI tuck-in, automatically parsing tracking emails from carriers and merchants into a unified delivery dashboard.
Why it matters
Apple Intelligence stumbled out of the gate over the past year. Siri's promised brain transplant was delayed in March, and notification summaries were notoriously buggy at launch. This WWDC feels like Apple quietly conceding ground on the fantasy of a fully autonomous digital assistant to focus entirely on utilities that actually work.
For users, baked-in phone translation and screen-aware lookups are immediately practical. You do not need to adopt an entirely new AI workflow; the features simply live where you already type, call, and screenshot. It makes the upcoming OS updates well worth an early trial for anyone dealing with multi-language calls or visual research.
For developers, free on-device inference is a genuine coup. Being able to run local generative tasks in three lines of Swift eliminates cloud inference costs and privacy compliance headaches for indie studios overnight. We expect a massive wave of third-party apps to adopt light text and quiz generation.
However, Apple is still playing catch-up to Google's Pixel ecosystem, which has offered screen parsing and call screening for years. Relying heavily on ChatGPT integrations for high-level reasoning also highlights what Apple’s own models still cannot handle alone. Shoring up daily utilities was a smart, necessary triage, but until Apple can ship a reliable next-generation Siri, its AI narrative remains visibly incomplete.
Sources (2)
- appleinsider.com(opens in a new tab)Originalappleinsider.com
- engadget.com(opens in a new tab)Reportengadget.com