AI
iOS
Apple
TLDR: LONDON—iOS 27 developer beta adds AI Clean Up, Extend, and Reframe in Photos, and the improved accuracy surprised a skeptic. Clean Up removed London passersby while Extend and Reframe changed framing with convincing results.
Key Takeaways:
- iOS 27 developer beta is early software for supported iPhones, with Apple Intelligence limiting devices to newer models
- Clean Up erases unwanted people or objects more seamlessly, Extend expands scenes with AI filled background, and Reframe shifts perspective
- The tools push Photos editing from simple touch ups toward controlled image transformation that can still feel believable
The funny part is not that Apple added AI photo tricks. It is that Clean Up actually behaves like a tool, not a gamble, making “cheating” feel optional.
The funny part is not that Apple added AI photo tricks. It is that Clean Up actually behaves like a tool, not a gamble, making “cheating” feel optional.
Q&A
If Clean Up now highlights multiple subjects, how will editors prevent accidental removal of important details like hands, signs, or pets?
Apple will likely lean on better subject detection and more granular selection so users can confirm exactly what gets erased.
What would make Extend feel better than a crop, beyond simply adding space?
Natural edge blending and consistent lighting that matches the original background would keep the added area from looking “generated.”
How do tools like Reframe change expectations for what a smartphone camera must capture in the moment?
They nudge users to trade perfect composition for capture speed, trusting AI to tidy perspective and framing later.
Why does the developer beta matter for trust, especially when AI edits can rewrite what a photo “means”?
Because early versions expose the failure modes, users can learn where the AI guesses incorrectly and adjust how they rely on it.
Could Apple Intelligence support wider iPhone models faster, and what would that shift in the user base?
If Apple expands device support, more people will experiment, which usually forces faster refinement as edge cases multiply in the wild.
No comments yet. Be the first to share your thoughts!