Sony is trying to get ahead of a growing backlash over how its “AI Camera Assistant” works on the Xperia 1 XIII, after a promotional post that was meant to show off the feature instead sparked mockery. The criticism wasn’t subtle: people pointed to example images that looked wildly unflattering, with some commenters arguing the assistant’s suggestions were less like helpful photography guidance and more like an automated path to bad results.
In response, Sony has moved to clarify what the assistant is actually doing—at least according to the company’s own description of the feature. The key point in Sony’s explanation is that the assistant isn’t presented as a traditional “editor” that directly rewrites your photo after you shoot. Instead, it’s framed as a suggestion engine that proposes different ways to adjust exposure, color, and background blur based on what the phone’s camera system detects in the scene.
That distinction matters, because it changes how you interpret the assistant’s behavior. If the assistant were simply applying edits behind the scenes, the debate would be about whether those edits are good or bad. But if it’s offering choices before you commit—essentially guiding you toward a better-looking capture—then the conversation shifts toward whether the assistant’s recommendations are grounded in real photographic judgment, and whether the examples Sony chose to demonstrate the feature are representative of what users should expect.
According to Sony, the assistant works by analyzing the scene you’re pointing the camera at. It looks at factors such as lighting conditions, depth information, and the subject in view. Then, rather than producing one “best” version, it generates four options. Those options are intended to represent different adjustments you can choose from, specifically around exposure and color, plus background blur—an area that often determines whether a portrait feels crisp and intentional or flat and distracting.
This “four options” approach is also where Sony’s messaging becomes important. The company’s product materials describe the assistant as giving users multiple directions, not forcing a single outcome. In theory, that means the user remains in control: you can pick the suggestion that matches your taste, your lighting preferences, or the mood you’re trying to create. In practice, however, the assistant’s value depends heavily on the quality of the suggestions it generates. If the four options are consistently unappealing—or if they skew toward a look that doesn’t match what many people consider natural—then the feature can feel less like assistance and more like a roulette wheel.
Sony’s clarification also touches on another phrase that has become part of the controversy: “the most photogenic angle.” In the company’s product video, Sony uses language that implies the assistant can help you find a better viewpoint. But the clip shown in that video appears to demonstrate something closer to zooming in than physically suggesting a new camera angle. That mismatch between wording and what viewers actually see is likely one reason the feature has been interpreted so negatively. People don’t just want better exposure; they want guidance that feels like a photographer’s eye—something that helps them move, reframe, or change perspective. If the assistant’s “angle” suggestion is really just a framing adjustment or a zoom recommendation, then the marketing implication may be stronger than the actual capability.
Sony’s latest explanation tries to narrow that gap by focusing on what the assistant does generate: suggestions tied to measurable scene characteristics. The company says it doesn’t edit photos, but provides options based on lighting, depth, and subject. That suggests the assistant is operating at the level of camera parameters and compositional effects rather than performing a full image transformation. In other words, it’s not described as replacing your photo with a new synthetic version or applying heavy-handed style changes. Instead, it’s positioned as a decision-support layer that helps you choose settings that will make the image look more polished.
Still, the public reaction indicates that the assistant’s output—at least in Sony’s posted examples—didn’t land well. When a demonstration shows results that look dramatically worse than what a typical user could achieve manually, skepticism follows quickly. Even if the assistant is technically “only suggesting,” the user experience is still judged by the end result. If the assistant’s suggestions are the ones being showcased, then the assistant is effectively being evaluated on those outcomes, regardless of whether Sony calls it editing or guidance.
There’s also a broader issue at play: how AI features are expected to behave in consumer photography. Many smartphone users have learned to treat computational photography as a spectrum. On one end are features that subtly enhance images—improving dynamic range, sharpening details, or stabilizing low-light shots. On the other end are features that aggressively reshape the look of a scene, sometimes producing artifacts, unnatural skin tones, or oversmoothed textures. Even when an AI feature is “just” adjusting exposure and blur, the aesthetic can still drift into the second category if the algorithm’s idea of “better” doesn’t match human expectations.
Sony’s assistant, as described, targets three levers that strongly influence perceived quality: exposure, color, and background blur. Each of these can easily go wrong. Exposure can be too bright or too dark, flattening contrast or blowing highlights. Color can shift in ways that make skin tones look off or make the scene feel artificially warm or cool. Background blur is especially sensitive: too much blur can look fake, while too little can make portraits feel cluttered. If the assistant’s four options consistently push these levers toward an unflattering aesthetic, users will interpret the feature as producing “bad photos,” even if the underlying mechanism is only parameter selection.
Sony’s response also implicitly raises a question about how the assistant is presented to users. If the assistant is meant to be a set of suggestions, then the interface and the timing matter. Does it show the options before capture, letting you choose the one you want? Or does it show options after capture, effectively turning the process into a post-shot selection? Sony’s explanation suggests the assistant is generating options based on what it sees, which could mean it’s integrated into the capture flow. But the controversy suggests that whatever the workflow is, the assistant’s recommendations are being experienced as a kind of automatic “look generator,” at least in the way Sony’s examples were perceived.
The company’s attempt to clarify “doesn’t edit photos” may also be aimed at a specific kind of criticism: the fear that the assistant is rewriting reality. Some users are wary of AI features that alter images in ways that are hard to understand or reverse. By emphasizing that it doesn’t edit photos, Sony is trying to reassure people that the assistant is not performing a hidden transformation. Instead, it’s offering choices that correspond to camera adjustments. That framing is meant to preserve trust: you’re not being tricked into accepting an AI makeover; you’re being offered a set of camera-based alternatives.
But trust is built not only on definitions—it’s built on consistency. If the assistant’s suggestions vary widely in quality, or if the best-looking option isn’t obvious, users will feel like they’re fighting the feature rather than using it. Four options can be helpful when at least one is clearly superior. It becomes frustrating when all four look wrong, or when the “best” option is subjective in a way that the assistant fails to capture.
Sony’s clarification comes after unwanted attention for the feature, and the timing suggests the company is responding to a specific wave of public scrutiny. The Verge reported on the issue, noting that Sony’s post demonstrating the assistant produced images that drew ridicule. Sony’s follow-up explanation is essentially an attempt to reset the narrative: the assistant isn’t supposed to be a photo editor; it’s supposed to be a guided suggestion system based on scene analysis.
There’s also an interesting tension between what Sony claims and what viewers infer from marketing clips. Product videos are designed to communicate capability quickly, often using simplified demonstrations. When a video says “photogenic angle,” viewers naturally assume the assistant will help them reposition the camera—move left, step closer, change height, or reframe. If the demonstration instead shows zooming, then the viewer’s mental model of the feature is different from the feature’s actual behavior. That difference can fuel disappointment, especially among people who already know that “zooming” is not the same as “finding the best angle.”
Sony’s explanation about lighting, depth, and subject is more concrete, but it also highlights the limits of what an assistant can do without understanding intent. Lighting and depth are measurable. Subject detection is also possible. But “photogenic” is a human concept. It depends on composition, storytelling, and context—things that aren’t fully captured by exposure and blur alone. An AI assistant can approximate what tends to look good, but it can’t always replicate the nuance of a photographer deciding what matters in a frame.
That’s why the assistant’s four-option design is both promising and risky. It acknowledges subjectivity by offering multiple variations. Yet it also assumes the assistant can reliably map scene characteristics to aesthetically pleasing outcomes. When that mapping fails—especially in the kinds of scenes used for marketing demos—the assistant’s credibility takes a hit.
Another angle worth considering is how these features are trained and tuned. Sony’s assistant is likely optimized for certain conditions: particular lighting ranges, common subjects, and typical smartphone viewing preferences. If the assistant is tuned for “safe” improvements—like slightly brighter exposure, a certain color bias, and a moderate blur effect—it might work well in many everyday situations. But if the marketing examples include edge cases—harsh lighting, mixed color temperatures, unusual backgrounds, or challenging depth cues—then the assistant’s suggestions could look dramatically worse than expected. That would explain why the assistant can be defended as “not editing” while still being criticized for producing unflattering results in specific demonstrations.
Sony’s clarification also suggests the company wants to steer the conversation away from accusations of direct AI editing and toward a more technical understanding of the feature. By describing the assistant as generating options for exposure, color, and background blur, Sony is essentially saying: this is computational photography assistance, not a generative image rewrite. That’s a meaningful boundary