Sony's AI Camera Assistant on the Xperia 1 VIII makes photos worse — and Sony is defending it
Sony's flagship Xperia 1 VIII ships with an AI photo assistant whose official demo images were so overexposed and overprocessed that rival executives called them engagement bait.
What matters
- Sony's AI Camera Assistant on the Xperia 1 VIII produces overexposed, overprocessed photos that look worse than the originals.
- Sony's own marketing comparison images were so bad that Nothing CEO Carl Pei suggested they might be engagement farming.
- The Verge confirmed after a week of testing that the feature is as bad as the demo images suggested.
- Sony is publicly defending the feature despite widespread mockery and a meme trend on X.
- The Xperia 1 VIII is otherwise an unusual flagship with a headphone jack, microSD slot, and thick bezels.
What happened
When Sony unveiled the Xperia 1 VIII in May 2026, it promoted a new feature called AI Camera Assistant, part of a broader "Xperia Intelligence" branding push. The feature works by presenting users with four automatically generated options that tweak exposure, color grading, and background blur after a photo is taken. Sony's marketing also claims the assistant can find the "most photogenic angle," though the demo only showed someone zooming into a static shot.
The problem: Sony's own before-and-after comparison images were startlingly bad. The "after" shots were vastly overexposed, with washed-out contrast, lost detail, and a cheap, overprocessed HDR look. A sandwich photo became a particular flashpoint — the AI-edited version looked like someone had cranked every slider to maximum. The Verge, after spending a week with the Xperia 1 VIII, confirmed that the feature is "exactly as bad as it looks."
Sony has since defended the feature, saying it works well and that the company doesn't understand the backlash. The Xperia 1 VIII itself is otherwise an unusual flagship: it includes a headphone jack, thick bezels, and a microSD slot — features most competitors have dropped.
Why it matters
This is a case study in AI features deployed without adequate quality thresholds. Sony is a company with deep imaging expertise and a respected camera division, yet its AI assistant produces results that look worse than no editing at all. That gap between brand promise and shipped output is exactly the kind of thing that erodes consumer trust in AI-branded features broadly.
It also illustrates a growing fatigue: when companies slap an "AI" label on a feature that visibly degrades the user experience, the backlash is swift and public. The incident became a meme, with users on X posting their own overexposed photos and joking that they had used "Sony AI Camera Assistant." For product teams, the lesson is straightforward — if your AI feature can't clear the bar of "better than doing nothing," it shouldn't ship under that banner.
Public reaction
Reaction was overwhelmingly negative and quickly turned mocking. Nothing CEO Carl Pei reposted Sony's comparison images and suggested the post might be "engagement farming" — implying Sony deliberately shared bad results to generate attention. Other users piled on, with a trend emerging on X where people deliberately overexposed their own photos and labeled them as Sony AI Camera Assistant edits. TechSpot noted that some observers initially thought the images were mislabeled because the quality drop was so extreme. Sony's decision to defend the feature rather than acknowledge the criticism only intensified the mockery.
What to watch
- Whether Sony issues a software update that meaningfully improves the AI Camera Assistant's output, or quietly walks back the marketing.
- How this affects Xperia 1 VIII reviews and sales, given that the phone's camera is a core selling point for Sony's enthusiast audience.
- Whether other smartphone makers recalibrate their own AI photo features in response to the backlash, or continue shipping aggressive computational processing.
- The broader question of whether consumer tolerance for AI-branded features that degrade quality is reaching a tipping point.
Sources
Public reaction
No Reddit discussion was captured for this story. Public reaction was concentrated on X, where users widely mocked Sony's AI Camera Assistant demo images and created a meme trend of deliberately overexposing photos. Nothing CEO Carl Pei publicly questioned whether Sony's post was engagement farming.
Signals
- Widespread mockery of AI-degraded image quality
- Skeicism that Sony's post was genuine vs. engagement bait
- Meme-driven backlash with users parodying the feature on X
- Industry competitor (Nothing CEO) publicly piling on
Open questions
- Will Sony acknowledge the criticism or continue defending the feature?
- Were the demo images representative of the final shipping software or an early build?
- Is there a software update path that could salvage the feature?
What to do next
Developers
If building AI-assisted image processing features, establish automated quality gates that compare AI output against the original and reject results that score worse on standard metrics like exposure, contrast, and detail retention.
Sony's failure shows that shipping AI output without a 'better than baseline' threshold can produce visibly worse results and public embarrassment.
Founders
Before launching any AI-branded consumer feature, run it past a panel of target users and compare AI-enhanced output side-by-side with unenhanced output. If users prefer the original, delay the launch.
The Xperia 1 VIII incident demonstrates that AI features that degrade the core experience damage brand credibility far beyond the feature itself.
PMs
Audit your product's AI features for cases where the AI output is worse than the non-AI default. Prioritize fixing or removing those before any marketing push.
Sony marketed a feature whose own demo images were mocked — a clear sign the feature was not ready for promotion and the marketing team lacked quality context.
Investors
When evaluating companies touting AI features in consumer hardware, scrutinize whether the AI demonstrably improves the user experience or is a marketing checkbox. Treat unproven AI claims as a reputational risk factor.
Sony's stock of imaging credibility did not protect it from backlash; AI features that visibly fail can erode trust in adjacent product lines.
Operators
If your company ships AI-enhanced outputs to customers (photos, text, recommendations), implement a rollback or opt-out path so users can always access the non-AI version.
The ability to revert to the original photo would have softened the Xperia 1 VIII backlash; giving users control over AI output reduces complaint volume and churn risk.
How to test
- 1Open the camera app on the Xperia 1 VIII and take a photo of a high-contrast scene (e.g., a meal on a table with mixed lighting).
- 2Trigger the AI Camera Assistant to generate the four suggested editing options.
- 3Compare each AI-generated option against the original unedited photo side by side.
- 4Repeat with a portrait subject and an outdoor landscape to test across scene types.
- 5Note whether any of the four options improve on the original in terms of exposure, color accuracy, and detail retention.
Caveats
- Results may vary by scene type and lighting conditions; test across multiple scenarios before drawing conclusions.
- Sony may issue software updates that change the AI Camera Assistant's behavior over time.
- The feature's 'photogenic angle' claim was not demonstrated in Sony's own marketing, so angle suggestions may be limited or absent.