Content authenticity

A Very Important Person made some waves in the content authenticity space recently.

Having developed Pixel’s strategy for Content Authenticity (with some very good help from my friend Sherif Hanna), and now having launched C2PA on every image and video created by the Pixel Camera app on Pixel 10, 9, and 8 devices, I couldn’t quite resist having something to say. I do mean launched – it was the rare time since transitioning to Leadership (ahem) that I got to build a real feature.

The key risk that a platform like Instagram faces is that, as the world inevitably changes more and more quickly, the platform fails to keep up. Looking forward to 2026, one new significant shift is that authenticity is becoming infinitely reproducible.

I think Adam here means mistaken-for-authenticity. I’ve been very careful with the word “authentic” since Florian Koenigsberger drew me into Real Tone in 2019. There is and isn’t* a definition of authenticity. What bothers users is the bait-and-switch of thinking it’s real and being wrong. Now, hm, if only I didn’t have to tap the three dots menu in Instagram and debate what’s supposed to be there, since “AI Info” only shows up if it is AI.*

Everything that made creators matter—the ability to be real, to connect, to have a voice that couldn’t be faked—is now suddenly accessible to anyone with the right tools. Deepfakes are getting better and better. AI is generating photographs and videos indistinguishable from captured media. The feeds are starting to fill up with synthetic everything. And in that world, here’s what I think happens.

Creators matter more

Power has been shifting from institutions to individuals for years because the internet made it so anyone with a compelling idea could find an audience. The cost of distributing information is essentially zero, which Ben Thomson has been pointing out as far back as 2014, so people can now bypass the traditional ways information spreads (newspapers on trucks, produced shows on television) and just go directly to an audience. We see it in everything from athletes who are more relevant than their teams to journalists who are more trusted than their publications.

What we’ve seen with the creator economy is individuals, not publishers, media companies or brands, establish that there is a significant market for content from people. Trust in institutions – government, media, corporations – has been declining for decades.In a world where we’ve all been inundated with heavily produced content from institutions we’ve turned to self-captured content from people we admire, creators.

We should all say thank-you, by the way, to Insta for creating much of the modern creator economy. They brought the friction of connecting to zero.

But we haven’t truly grappled with synthetic content yet. We are now seeing an abundance of AI generated content, and there will be much more content created by AI than captured by traditional means in a few years time. We like to talk about “AI slop,” but there is a lot of amazing AI content that thankfully lacks the disturbing properties of twisted limbs and absent physics. Even the quality AI content has a look though: it tends to feel fabricated somehow. The imagery today is too slick, people’s skin is too smooth. That will change; we are going to start to see more and more realistic AI content.

Adam is making the classic mistake here of assuming his user catches subtleties – The everyday user literally already cannot tell, and has not been able to tell for the last couple years already, what’s generated by AI and what’s created by a camera. They only ask if it’s AI if they already have a reason to disbelieve it, and you can imagine that is all tied up in how they identify and what their in-groups are.

Authenticity is fast becoming a scarce resource, which will in turn drive more demand for creator content, not less. The creators who succeed will be those who figure out how to maintain their authenticity whether or not they adopt new technologies. That’s harder now—not easier—because everyone can simulate authenticity. The bar is going to shift from “can you create?” to “can you make something that only you could create?” That’s the new gate.

I think it’s true that real people will always win. Much like the discipline of “art” changed from sculpture to painting to photography to video creation, new tools don’t erase humanity or artistry. What underlies the tools will always be the human connection. We’ll see a lot of early-stage creators competing against models to climb to the next level, but the most premium will be person-led. GenZ photo dumps, the resurgence of film photography even among Millennials, surging sale prices of “digicams”, and the new Luddite’s* newfound love of typos (because AI doesn’t make typos!) proves it. The backlash against AI only serves to premium-ize humans.

The raw aesthetic

Just as AI makes polish cheap, phone cameras have made professional-looking imagery ubiquitous—both trends cheapen the aesthetic.

Unless you’re under 25 and use Instagram, you probably think of the app as a feed of square photos. The aesthetic is polished: lots of make up, skin smoothing, high contrast photography, beautiful landscapes.

That feed is dead. People largely stopped sharing personal moments to feed years ago. Stories are alive and well as they provide a less pressurized way to share with your followers, but the primary way people share, even photos and videos, is in DMs. That content is unpolished; it’s blurry photos and shaky videos of people’s daily experiences. Think shoe shots and unflattering candids.

That goal is dead, but GenZ’s photo dumps are just as curated as Millennials’ feeds were, just curated for a different goal. It’s still branded, just the brand goal is different. It’s deeply embedded in us to manage the image we project. Even “candids” are posed, and there’s plenty of content on Reels teaching people to do that better.

This raw aesthetic has bled into the zeitgeist of public content and across art forms. Think @jordan_the_stallion8 shooting videos in a bathroom mirror, or @pitchfork’s description of @mrcameron_winter’s voice as “a slurred, straining warble.”

The camera companies are betting on the wrong aesthetic. They’re competing to make everyone look like a professional photographer from the past. Every year we see phone cameras boast about more megapixels and image processing. We are romanticising the past. Portrait mode is artificially blurring the background of a photograph to reproduce the soft glow you get from the shallow depth of field of a fixed lens. It looks good, and we like to look good.

I’ll forgive Adam for this* – There’s objective quality (SNR, MTF, aliasing), subjective quality (color target, global tone-mapping, sharpening), and there’s marketing quality (most megapixels, longest zoom, biggest sensor). Adam’s complaint is with objectives and marketing, but those are essentially orthogonal to subjective qualities. I can make a DSLR image look just as “overprocessed” as the worst smartphone’s* with a couple sliders in Lightroom (ahem, Clarity, Dehaze). Smartphone cameras produce incredibly higher quality now than they did just a few years ago (especially objective and marketing), but it’s hard for subjective quality to keep up with the latest fads – especially when real feedback is masked by the availability of apps that make applying whatever filter you want as easy as one tap.

Gee, wouldn’t it help build a better camera if we could see what people were doing in their editors to make their images “better” before sharing? Say, an app where people share tons of DMs and posts every day?*

But flattering imagery is cheap to produce and boring to consume. People want content that feels real. We are going to see a significant acceleration of a more raw aesthetic over the next few years. Savvy creators are going to lean into explicitly unproduced and unflattering images of themselves. In a world where everything can be perfected, imperfection becomes a signal. Rawness isn’t just aesthetic preference anymore—it’s proof. It’s defensive. A way of saying: this is real because it’s imperfect.

Defaulting to skepticism

Relatively quickly we are going to see the AI tools that create content mature and the range of aesthetics that they can produce expand. We’ll go from the Midjourney realistic video game aesthetic and imitating Wes Anderson and Studio Ghibli films to being able to direct an AI to create any aesthetic you like, including an imperfect one that presents as authentic.

So funny how Adam thinks this hasn’t already happened. My friend, you are a couple years behind!

At this point we’ll need to shift our focus to who says something instead of what is being said.

False – We’ll shift our focus simply to whether the person who said something was a person or not. I believe humanity alone will have value.

For most of my life I could safely assume that the vast majority of photographs or videos that I see are largely accurate captures of moments that happened in real life. This is clearly no longer the case and it’s going to take us, as people, years to adapt.

I said it on Tyler Stalman’s video podcast with Sherif and I’ll say it again. When I was in middle school, I learned that I should trust .gov more than .com; kids today need to learn in middle school that AI can impersonate their own mother on the phone. This is a basic digital literacy task.

Over time we are going to move from assuming what we see is real by default, to starting with skepticism when we see media, and paying much more attention to who is sharing something and why they might be sharing it. This is going to be incredibly uncomfortable for all of us because we’re genetically predisposed to believing our eyes. Malcolm Gladwell, in his book Talking to Strangers, articulated elegantly that we,if there’s as a species, default to truth because the evolutionary and social benefits of efficient communication and cooperation far outweigh the occasional cost of being deceived.

Adam says “incredibly uncomfortable”, and I say “never going to happen”. Or perhaps Adam thinks of generational change, and I think of change I can make for the people around me right now. We are never going to be able to insert an “is this real?” loop in the middle of doom-scrolling Reels without the platform’s help.

Social media platforms are going to come under increasing pressure to identify and label AI-generated content as such. All the major platforms will do good work identifying AI content, but they will get worse at it over time as AI gets better at imitating reality. There is already a growing number of people who believe, as I do, that it will be more practical to fingerprint real media than fake media. Camera manufacturers could cryptographically sign images at capture, creating a chain of custody.

Count me among those who believe the same! Not only could cameras sign images cryptographically, some of us do! (Now available for photos and videos captured with Pixel Camera app on Pixel 10, 9, and 8 series.)

It’s so, so much worse than Adam shares, though, to label only AI content. Imagine you have a product that applies a visible label to all AI content and no label to the rest. Users scroll and scroll, and eventually you stop one and ask “is this AI or not?” The overwhelming majority of users say “no label, so it’s not AI”. But that’s not right – the truth is you don’t know anything. It could be from a camera and unlabeled; from AI and unlabeled; it could be from AI with a label that a malicious poster stripped out.

And when it is labeled AI, and you ask the user to judge how misleading the content is, they’ll give it a lower score than all those “not AI” ones. But in fact the content that’s willing to self-identify with the label as AI isn’t nearly as concerning as the master manipulator who strips the label intentionally before posting.

So, by labeling only AI content as such and leaving the rest unlabeled, we’ve increased trust in the least trustworthy content. It’s so, so much worse than just difficult to label only AI – it’s harder to do.

Also, we have tons of robust technology (SSL, HTTPS, TLS, symmetric-key, asymmetric-key, etc.) to verify information when that information has value to both parties (“if you give me your password, I’ll let you read your email”), but we have no robust technology (steganography) to verify information when that information is harmful to one party (“This content is generated with AI”). Labeling AI just pushes technology – not just people – out of their comfort zones.

Labeling content as authentic or AI-generated is only part of the solution though. We, as an industry, are going to need to surface much more context about not only the media on our platforms, but the accounts that are sharing it in order for people to be able to make informed decisions about what to believe. Where is the account? When was it created? What else have they posted?

So what?

In a world of infinite abundance and infinite doubt, the creators who can maintain trust and signal authenticity—by being real, transparent, and consistent—will stand out.

As for Instagram, we’re going to have to evolve in a number of ways, and fast. We need to build the best creative tools, AI-driven and traditional, for creators so that they can compete with content fully created by AI. We need to label AI-generated content clearly, and work with manufacturers to verify authenticity at capture—fingerprinting real media, not just chasing fake. We need to surface credibility signals about who’s posting so people can decide who to trust. And we’re going to need to continue to improve ranking for originality, but tackling algorithmic transparency and control is probably best left for another essay.

This is one of those classic two-faced problems. Platforms won’t show the metadata until cameras capture it, and cameras won’t capture it unless platforms show it. I can point to at least two key ways Pixel’s C2PA implementation could improve if platforms were committed to showing all the relevant context about each image and video.

Glad to be on the cutting edge here at Pixel Camera! And very glad to have such a close relationship with Insta that I know personally the folks working on implementing whatever “evolve in a number of ways” means. Here’s to supporting the unfailing premiumness of human creation.

• A story for another time, I think.
• This isn’t the first time I’ve taken a potshot at an Instagram exec – I also did on The Verge in 2019. Sorry! I’m easy to through official channels if you want to chat!
• I do mean the negative connotation; those who oppose AI globally oppose obvious progress and the further democratization of creation. But we can recognize both those Luddites and the new ones do have some very good points!
• Again, I’m happy to chat. I have some advice for improving your Instagram camera app!
• I won’t name names!
• Ok, ok, I promise it’s the last potshot at Insta. Maybe?
• Still a topic for another time.