Apple Photos can find any photo, but it has no idea what photos are for

On wicked problems, emotional design, and why more AI isn’t the answer

Open Apple Photos right now. Try searching “birthday” or “dog”. It finds what you’re looking for in under a second. Look back two years and behold — there’s an entire cinematic slideshow waiting for you complete with music, animations and a sophisticated header and whatnot. All of it done without you asking. Without you even knowing it was happening.

And let’s be honest — it’s impressive. Yet, for many of us, myself included — something feels off. True, all of the photos are there. Everything opens up fast, search works impeccably, the interface is well thought out. But the experience of looking at the pictures, of feeling something when looking at them is somehow absent. You scroll through thousands of pictures without really stopping and the app just doesn’t care.

That gap — between what the product can do and what it actually feels like to use — has a name. It’s not a technical failure. It’s a design thinking one. Apple came up with a genuinely impressive solution to a problem that it defined on its own terms without ever bothering to ask what problem people have with their photos.

But before we dive any further…

Two frameworks that you should know about

In 1973, Horst Rittel and Melvin Webber published a paper that shifted the way designers approach problems. The two were urban planners who sought to understand why social issues such as housing, poverty, public transportation, etc. resisted all attempts at solving them in a neat logical manner. Their argument is that there is a difference between “tame” problems and “wicked” ones.

But what exactly are “tame” problems? The kind that are easily defined. They can be described in a concrete manner, you can test your solution against them and you know when you have solved them. Wicked problems, on the other hand are different in almost every way. There is no single formula to solve them. How you frame the problem determines what solutions will look like, which means two people looking at the same situation can arrive at entirely different outcomes. There’s no stopping rule, no moment where you can say with confidence that you’re finished. And every attempt to solve them has consequences.

As stated by Rittel and Webber,

“every implemented solution to a wicked problem will generate waves of consequences over an extended period of time”.

They argue it becomes morally unacceptable for planners to treat a wicked problem as though it were tame. Solving a messy human problem by applying logical thinking does not solve the problem but makes it even worse.

In the context of Apple Photos, Tim Brown’s “design thinking” provides a methodology for approaching the problem. Brown believes that the starting point of any design process is always the person. Not the technology, not the business model — the person. Design thinking is a discipline that tries to match what people actually need with what’s technologically possible, and that match only happens if you understand the need first — through observations and empathic research of users’ needs. The empathy phase — the slow, unglamorous, easy to skip — is what decides whether a product ends up being truly strategic or just tactical. A tactical product only refines an already formed idea. A strategic one figures out what to build before it starts building.

So far, Apple Photos has the technology side of that equation handled rather well, but not the other half — understanding what people actually need.

But then, what has Apple done right?

Let’s be objective here. Some of the design solutions in Apple Photos are genuinely thoughtful.

The biggest shift Apple made was moving away from folder-based organisation, where users were expected to sort everything manually — towards a time and location-based library. This was absolutely the right call because Apple designed around actual behaviour instead of trying to design something in an ideal way. People simply take too many photos to sort them and they give up doing that sooner or later (unless that’s your jam, like yours truly). The design solutions that work around this reality are exactly what you need to create a useful product.

The other feature that is also well-designed, well-thought out is search. It works on a command model — the user defines what they want, the app retrieves it. Type “cat” or “Europe 2024” and it’s there immediately. This turns an overwhelming archive into something navigable, on the user’s own terms.

But there’s one thing to note here. What Apple was doing in both cases was taking a wicked problem — ‘how do people manage a growing visual record of their lives?’ — and reducing it to tame, solvable sub-problems. Navigation. Retrieval. Organisation. It is an all too familiar strategy, and it works. The trouble is, it only holds within the tame layer. The moment Apple tries to go further— to address not just how people store photographs, but how they feel about them — the strategy stops working.

When did it become unmanageable?

It happened with the introduction of the Memories feature, launched in 2016 with iOS 10. This function automatically creates a slideshow of photos in your photo library. Selecting images, adding music, applying transitions, presenting the result as a curated emotional retrospective. One Year Ago Today. Best of Summer. Your Trip to Berlin. Technically brilliant, yet for a lot of people, genuinely disorienting.

Here’s why…

The question Apple was trying to answer — ‘how should a person relate to their past photographs?’ — is not a tame problem. It’s a wicked one. There’s no single right answer. It depends entirely on who you are, what that period of your life meant to you, whether the people in those photos are still around and in what capacity. Would any reasonable person see a slideshow of a difficult time as a gift? Not really. The algorithm can’t distinguish between situations because the algorithm operates on the assumption that people want to be reminded of their good times.

The consequences of that assumption have been documented. In 2025, in a study conducted by the University of Alberta, 15 people who had been through gender-based violence were interviewed using a trauma-informed approach. Every one of the interviewed people had an intense reaction of panic, nausea, physical symptoms when the Memories surfaced photos of abusive ex-partners. In some instances, the shutdown lasted for several days. One of the participants reported going into full panic. after Memories showed a picture of her abuser with cheerful and upbeat music playing in the background. It was the music, in particular, that survivors just could not stand. This detail is impossible to anticipate even for an advanced algorithm. But it’s exactly the kind of detail an empathy-led process — one that genuinely asked who uses this feature, and under what circumstances — might have caught.

A related issue surfaced in April 2022, when journalists at 9to5Mac discovered Apple had quietly updated iOS to prevent photographs taken at Holocaust memorial sites from appearing in Memories. This move revealed something telling: Apple’s developers already understood that the algorithm was not capable of understanding the emotional context, but instead of fixing this flaw in the technology, their solution was to compile a manual list of such sensitive places. But that decision created new problems immediately: who defines whether certain places are sensitive?Moreover, why are only those particular places sensitive and not others? Or why is it not a hospital, where someone has passed away, or a house, which is the place of a terrible breakup? It is impossible to compile an exhaustive list in this situation. This is a near-perfect illustration of what Rittel and Webber described — every solution to a wicked problem leaves traces that cannot be undone.

There is one more aspect of Memories that is hardly ever discussed in public: Apple (like Google Photos) has a commercial interest in collecting your pictures and analysing them. All the photos that you take, all the faces that you recognise, and all the moments that the app decides are significant for you become training materials for AI models. A design process that is truly based on users’ needs should have shown that side of the feature to users and allowed them to manage it somehow.

This pattern became most visible with the iOS 18 overhaul in late 2024. Apple called it “the biggest-ever” update to Photos. They introduced an all-new unified scrollable interface, smart collections powered by AI, Memory Maker feature. Within days, users from all forums and social media platforms were referring to it as confusing, overwhelming, and unusable. Complaints included a cluttered interface that made finding specific photos difficult, redundant auto-generated albums, missing features people had quietly relied on for years, and AI curation crowding out basic organisation. Apple had to roll back many of the changes in iOS 18.2 just three months later. And by 2025, with iOS 26, brought back the tabbed interface entirely. Craig Federighi’s acknowledgment at WWDC was as close as Apple gets to admitting a mistake: “Many of you missed using tabs in the Photos app.”

Rittel and Webber could not have scripted it better -

A tame solution, applied to a wicked problem, producing a fresh wave of consequences.

The deeper design failure

Let’s consider the problem from Brown’s perspective — the failure is methodological before it is anything else. Tim Brown says that true innovation requires a “thorough understanding, through direct observation of what people want and need in their lives.” Not what their data suggests. Not an understanding of the way they interact with the app. An understanding of what they really need on a level that goes beyond what any usage metrics show.

If Apple had genuinely observed the way people interact with their photographs — not just how they interact with photo apps, but the basic human reality underneath — they would’ve seen something worth sitting with.

Think about how photographs lived in homes before smartphones. The box of prints your grandmother kept in a drawer. The album someone would pull out and pass around at a family reunion. That one framed photograph on the desk picked out from hundreds of others. All of these contained intentionality. Someone chose those photographs. They meant something before they were put on display.

Apple Photos makes those choices for you. And I feel that’s exactly where the emotional gap comes from. Curation without intention means curation without you. That’s the tactical Vs strategic difference mentioned by Brown. Apple Photos is a very tactical product presented as a strategic one. It is superb at presenting photographs but less so at understanding what they are for.

What better would look like?

We have already established that — if memory is a wicked problem, it can’t be solved in the way tame problems are solved. There’s no perfect version of the app that “gets it right”. But that’s not the same as saying nothing can be done. The approach has to change from solving to re-solving; that is, iterating thoughtfully, with humility, and knowing that any proposed solution is going to produce its own new consequences.

…What would that actually look like?

I believe it starts with asking a different question. Instead of ‘how do we surface the best pictures’, the better question is: ‘what role do photographs play in people’s lives? And the answer, which is based on direct observations rather than the behavioural data alone, would probably look less like automation and more like structure for intention. Applications that prompt you to choose, rather than choosing for you. An interface that says: here are the photos from this month — which ones do you want to keep? Not because an algorithm decided they were important. Because you did.

The trauma research points the same way. According to Little and Slane, there is a need for a trauma-informed approach to designing digital experiences where people get agency before they’re harmed, not after.

At its core, it’s another design thinking exercise -

Understand who your users actually are, including the most vulnerable among them, before you build features that reach into the most sensitive parts of their lives. This also means building for the social side of photo sharing, which apps ignore completely nowadays. Photographs historically haven’t been just a private archive — they are shared objects. You show them to people. You print them and give them away. You sit with someone and look at them together. A Photo app that considers these facts asks: Who do you want to share this with? Instead of making a Memory for people to watch alone with music.

This isn’t uniquely an Apple problem. It’s what happens the moment any design team becomes more fluent in its own technical capabilities than in the lives of the people it’s building for. That’s the moment a product stops being human-centred, even when it still looks that way on the surface.

Photographs hold something that’s hard to name but easy to feel. They’re how we hold on to people and moments that no longer exist in the present. A design that understands that will not try to do that work for you.

It would help you do it yourself.

References

Brown, Tim. 2008. “Design Thinking.” Harvard Business Review, June 2008, 1–9. Reprinted in Developments in Design Methodology, edited by Nigel Cross, 167–177.

Clover, Juli. 2023. “iOS 17 Photos Features: Pet Recognition, Recipe Identification, One-Tap Crop and More.” MacRumors, September 15, 2023. https://www.macrumors.com/guide/ios-17-photos-and-camera/

Hoskins, Andrew. 2024. “Between Automated Memory and History: Blocking ‘Sensitive Locations’ from Apple Memories.” Memory, Mind & Media 3. Cambridge University Press. https://doi.org/10.1017/mem.2024.3.

Little, Olivia, and Sara Slane. 2025. “Features like iPhone’s and Facebook’s ‘Memories’ Can Retraumatize Survivors of Abuse.” The Conversation, February 2025. https://theconversation.com/features-like-iphones-and-facebooks-memories-can-retraumatize-survivors-of-abuse-231897.

MacRumors. 2024. “iOS 18 Photos App Redesign: Two Months Later, Users Still Divided.” MacRumors, November 21, 2024. https://www.macrumors.com/2024/11/21/apples-photos-app-overhaul-controversial/.

r/iphone. 2024. “Photos App on iOS 17 / Photos App on iOS 18.” Reddit, December 2024.

https://www.reddit.com/r/iphone/comments/1h442zg/photos_app_on_ios_17_photos_app_on_ios_18/.

Rittel, Horst W. J., and Melvin M. Webber. 1973. “Dilemmas in a General Theory of Planning.” Policy Sciences 4 (2): 155–169. Reprinted in Developments in Design Methodology, edited by Nigel Cross, 135–144.

Apple Photos can find any photo, but it has no idea what photos are for was originally published in Bootcamp on Medium, where people are continuing the conversation by highlighting and responding to this story.