I don’t believe there’s any legit demand for people wanting to generate sloppified fictions of friends & loved ones. By Jared White #
In October of 2024, Apple software head Craig Federighi was asked by Wall Street Journal reporter Joanna Stern why Apple wasn't going as far as other tech companies in the area of “AI photo generation” and this was Craig's reply:
I would say even the ability to remove that water bottle is one that there were a lot of debates internally. Do we want to make it easy to remove that water bottle or that mic? Because that water bottle was there. The demand for people to want to clean up what seemed like extraneous details to the photo that don't fundamentally change the meaning of what happened has been very, very high. And so, we were willing to take that small step.
But we are concerned that there's a great history to photography and how people view photographic content as something they can rely on is indicative of reality. And our products, our phones are used a lot, and it's important to us that we help purvey accurate information, not fantasy.
(h/t: Greg’s Gadgets for surfacing this clip) So…what has changed between October 2024 and June 2026?
Other than the relentless drumbeat of “wHy iS AppLe sO fAr beHiNd in Ai??” in tech business rags, nothing!
Why has Apple taken a sickening lurch in a totally different direction, contradicting their previous statements, with a spate of “AI photo generation” (aka slop) features in the OS 27 release cycle?
Were there no more debates internally?
Did the slop machinists officially win the debate for good? Are the dissenting voices no longer in charge? Is Craig Federighi totally cool with Apple’s new direction? Or was he told by his boss (T. Cook) to make it happen anyway?
Is there no longer “a great history to photography” and do people not care anymore that photographic content is “indicative of reality” by and large?
Is it no longer important for Apple to help purvey accurate information, not fantasy?
These are troubling questions indeed.
But it gets worse.
Are you revisiting the role computer technology should (and should not!) play in our lives? Stay in the loop with my newsletterCycles Hyped No More:
Your Likeness is Not Your Own #
Friends and strangers alike can now completely manipulate your appearance and toss you about a virtual scene like a rag doll, at whim, whenever they choose.
And there’s nothing you can do about it.
In fact, Big Tech moguls are out there specifically campaigning against deepfake legislation, because something something free speech!
You have no control whatsoever over what people are doing to your likeness. And sure, this is something that’s been vaguely true ever since the arrival of Adobe Photoshop. Some might argue that photographic manipulation is as old as the medium itself, and our comfort that “a photo depicts reality” has always been a sort of fiction.
But it’s undeniable that Generative AI has completely changed the game when it comes to photographic manipulation, allowing you to produce at scale completely new scenes showcasing real-world people—an icky sloppy version of those people, yes, but to the undiscerning eye, appearing “authentic”.
Until this month, Apple had been a bit of a lone holdout among large tech companies. Sure, we’d been handed the Image Playgrounds app as part of the initial rollout of Apple Intelligence, but none of those images could be deemed “photorealistic”…in fact if you weren’t feeling charitable, you’d call them just plain weird.
Now Apple has mostly caught up to the mainstream of the slop machines, and instead of refraining from producing images of people with the proper safeguards—or requiring verified opt-in from the people involved via face detection—Apple has decided to throw in with the other Big AI boys and go whole-hog for the slop.
Ethics be damned!
Apple is Providing Everyone The Tools to Harm You Directly #
As I emphasized in my recent Vibe Coded episode, it is a wild breach of morality for Apple to hand (in theory) every one of their users a deepfake image generator. There are so many ways this could go horribly, horribly wrong.
Look, it’s bad enough that features like “Reframe” and “Extend” will outright just generate non-existent slop people, unreal places, and fictional things, and still let you call that a photo. But at least you can argue the utility of some of that. Instead of spending hours or days in Photoshop as a graphics design expert to make those types of corrections, “AI can do it for you.” As Craig stated in his 2024 interview, “the demand for people to want to clean up what seemed like extraneous details to the photo that don't fundamentally change the meaning of what happened has been very, very high.”
I believe that.
What I don’t at all believe is that there’s demand for people wanting to generate sloppified fictions of friends and loved ones in any meaningful way.
At best, it’s a cringe gimmick…another variation on the Shrimp Jesus phenomenon. Brainrot.
But at worst, this is a feature ripe for abuse, able to be weaponized in increasingly disturbing ways. And even when it’s not used by a bad actor, it can still be abusive and harmful.
This research paper from all the way back in 2023 highlights a wide variety of potential biases prevalent in generated imagery. Here is a helpful chart to kick off the discussion:
When you reach for a photo of a person, even a friend, in your photo library, and ask the AI image generator to place that person in a new setting, with new clothes, accomplishing any imaginable task, the likelihood that these biases come creeping in is quite high.
I’m simply waiting for the news articles to come rolling in of how a wheelchair user was sent a “photo” of themselves walking around, or lost limbs reappearing, or identity-bound facial or body features disappearing. (What happened to my tattoo?!)
Unfortunately, such news is often buried in favor of the more lurid abuses out there such as Deepfake NCII (Non-Consensual Intimate Imagery). We’re all aware of the horrors of Grok, and I’m assuming that Apple has had least gone to painstaking lengths to forbid users from using their new tools to perpetuate such abuse. Still, that doesn’t mean it’s impossible. And the milder forms of this are still quite bad. “Whoa look at that, Apple let me put you in a bikini!” is absolutely within the realm of possibility. Heck, Apple’s own demo showed a woman being slightly dressed down…a real person previously wearing a thick coat is suddenly placed in a scene where they’re “wearing” a short-sleeve t-shirt. I know for a fact there are women who do NOT want to show their bare arms to anyone in public. Who’s providing any consent for this type of manipulation?
Computers Don’t Have Opinions #
Of less concern perhaps, but still very much of concern to me, is the conversational nature of what Apple showed off with the new Siri AI. This screenshot is maximum cringe, as well as totally deceptive:
Student:How does this sound?Siri:It sounds great! You have a really engaging opening that paints a vivid picture, and your points about sustainability and education are well-reasoned. The humor at the end adds a nice touch of personality.
Liar! You lie!
And to be crystal clear, I’m not saying Siri is lying. That would be silly, because Siri is incapable of telling the truth either. Siri, in fact, has no opinion on if the student’s speech sounds great, good, mid, or bad, because Siri is incapable of having any opinion on the matter!
Computers don’t have opinions.
No, Siri isn’t lying here. Apple is lying. Just like all other unethical corporations in the Big AI space, Apple has created a sycophantic chatbot which glazes users, providing an utterly false impression that there’s any value whatsoever to the assessment the chatbot has made. There is none. It is a valueless exchange.
Siri could have easily responded “It sounds terrible. Your writing stinks. You’re a goddamn asshole and I hate you and I want you to go crawl into a hole and die, loser!” and that would have meant the exact same thing. That is to say, nothing at all.
Siri is incapable of producing meaning, only synthetic text. Computer strings to be precise, UTF-8 encoded. The only “meaning” in any of these outputs are what WE as humans imbue when we read them. It’s a bit like the dog owner who “talks” for their dog. Their dog pants and “smiles” and goes bow-wow and the human interprets that into a whole conversation. And sure, it’s pretty cute—when we’re talking about dogs! When we’re talking about chatbots however, it’s weird and gross.
Stop anthropomorphizing chatbots. These are not human-like entities under any circumstances. They are dataset extruders, nothing more. As Emily M. Bender and Nanna Inie write:
“AI” is not your friend. Nor is it an intelligent tutor, an empathetic ear, or a helpful assistant. It can not “make up” facts, and it does not make “mistakes”. It does not actually answer your questions. Such anthropomorphizing language, however, permeates the public discussion of so-called artificial intelligence technologies. The problem with anthropomorphic descriptions is that they risk masking important limitations of probabilistic automation systems, which make them fundamentally different from human cognition.
The people and companies selling “AI” technologies routinely use language that portrays their systems as human-like — “reasoning capabilities”, “hallucinating”, and artificial “intelligence”. The media has largely let them set the terms of the debate, right down to the terminology used in any discussion of these systems. But even the most flawless execution of a task typically associated with intelligence does not make a system “intelligent”, and the framing of systems as humans or human-like is misleading at best, deadly at worst.
Shame on Apple for showing so little leadership in this area, sacrificing true UX in the name of market expediency. Any answer to “How does this sound?” which isn’t politely declining to answer (e.g., “Apologies, I am but a mere chatbot lacking the ability to make value judgements. I can only provide a statistical analysis of your word and sentence structure.”) is deceptive and potentially cognitively damaging—a first step towards chatbot psychosis.
Computers don’t have opinions. Humans do.
The John Ternus of It All #
We don’t yet know how incoming Apple CEO John Ternus really feels about all this. We do know that outgoing Apple CEO Tim Cooks has very little taste. He’s never been a product guy. He has cared about certain admirable things outside of product design in the past…arguably he cared much more about charity than Steve Jobs ever did. But trusting a character like Cook to walk into an early product demo and shout “what the FUCK is this hot garbage?!” is like wishing an accountant would take a really strong stance on cinematography.
Who at Apple has had or will have the ability to see a meeting agenda with the headline “deekfake AI image generator” and say “Sorry, this just isn’t Apple. We need to kill this feature. Do better.” ??
I hope and pray that person is John Ternus—or at the very least, his product design instincts will lead him to putting the right people who place who do have that authority.
Because, if not, we are completely doomed. There is no other Big Tech company left standing, fighting the good fight against the worst excesses of the AI hype bubble. If we can’t trust Apple in this, we can trust no one.
Linux nerd on Mastodon:Finally Jared! You’re finally getting it!
OK, OK. Fine. You win. 😂
But seriously, in my optimism I do believe WWDC26 will be the last time we see this sort of weird-ass product cycle. Apple Unintelligence, as I like to call it, seems to have reached a place where most pundits are saying “it’s good enough” as compared with Google etc. My take is that the AI bubble will have fully burst before WWDC27, with hyperscalers in deep, deep financial trouble and markets trying desperately to course-correct and minimize AI-driven loses. That will give John Ternus and a refreshed management layer cover to de-emphasize AI and focus on what users really want from Apple.
As long as Siri “works” for the trivial things people do rely on chatbots for, hopefully a lot of the other stuff can be quietly walked back. No, I don’t want to see Mawmaw in a punk outfit riding a skateboard, and trust me, no one else does either!
Judging from all the backlash I’ve seen even in typically sycophantic Apple-loving press outlets, the slop generation features are by far the most egregious.
They’ll be the first to go. 🤞
💬 Discuss this essay in the Human Web Collective Discord ➡️ Thank you for reading Cycles Hyped No More. Join Intuitive+* *if you haven’t already to support my independent publishing, and please share this newsletter with a friend! See you here next week.
Jared ✌️ 🤔🌩️ Things that make you think:
“Backrooms” director Kane Parsons recently sounded off on AI in filmmaking during a recent interview with The Australian. The 20-year-old filmmaker said that he was “in the same boat as most well-adjusted people,” and does not want to see the technology take over Hollywood.
“If I could snap my fingers and make generative AI disappear forever, I probably would,” Parsons said. “Creatively, I get no enjoyment from using those tools. It defeats the purpose entirely for me.”
–‘Backrooms’ Director Kane Parsons Says Using AI ‘Defeats the Purpose’ of Filmmaking (Variety)