Constructing Reality, or Breaking the Rules?

My fiancée and I watch very little live television. One major exception, however, is the MLB postseason—provided that either the Phillies, New York Mets, or Washington Nationals have made it to the playoffs. Truth be told, this doesn’t happen all that frequently, so even then we watch very little live TV.

But when the stars do line up just right, not only do we get to experience the excitement (yes, I said it!) of multiple games of baseball; we also get to feast our eyes, ears, and brains on that most surreal form of secondary entertainment known as commercials.

Two things stood out as I watched this year’s fare: (1) Musical numbers (some knowingly tongue-in-cheek and others missed-the-memo earnest) are apparently in vogue, and (2) smartphones—and in particular, smartphone cameras—are becoming increasingly creepy.

I’m speaking, of course, about the Pixel 8 series, which Google launched in early October, just as the Wild Card round was getting started. As the commercials were happy to inform us—backed up by cool, techno vibes—these phones are packed full of AI-enhanced photo editing features.

Cause for Alarm?

Now, I should pause for a moment and note that artificial intelligence and machine learning have played a role in digital image processing for some time now. Apple, for instance, uses “Deep Fusion” to optimize photos; Canon does something similar with “deep learning.” For the most part, we don’t dwell on the calculations our phones and cameras perform under the hood. But the thing that kept freaking me out, commercial break after commercial break, was a new user-facing feature that Google has named “Best Take.”

As demonstrated in the advertisements, the feature allows you to take several photos in short succession and then create one “perfect” image by choosing the best face for each subject. It’s a great way to correct for family members that won’t stay still or that one friend who was blinking in every photo except for the one where you’re mid-sneeze. Instagram never looked so good!

The Washington Post’s Geoffrey A. Fowler sums up my discomfort when he asks, “Is this a line we want to cross?” He explains the deeper issue like this:

As much as I enjoyed using [Best Take], there’s an uneasy casualness about letting AI edit the faces in the smartphone photos we rely on to archive our memories. It’s allowing AI to help standardize ideas about what happiness looks like—an escalation of the cultural pressure we’ve been grappling with on social media to curate smiling faces and perfect places that don’t always reflect reality.

The column goes on to grapple, somewhat ambivalently, with the question of “How fake is too fake?” Like me, Fowler considers the role that software and algorithms have long played in photography, but he also raises fair concerns about slippery slopes and sexism, ultimately wondering aloud what the purpose of a photograph should even be. Is it a “record of a moment,” as Fowler puts it, or something else?

The Classroom Perspective

As someone who has taught photojournalism to high school students for around 13 years, I could not help but let this question sit with me for a while. Why do we create and share photos, let alone publish them in newspapers?

When I pose this question to my students, the consensus they usually settle on is that photos allow us to capture the reality of a particular point in time in a way that feels somehow “more real” or, at the very least, augments the written/spoken word.

We’ll look at Pulitzer Prize-winning photos together—moments of victory, tragedy, and everything in between—and talk about how the angle, the framing, the lighting, or even just being in the right place at the right time and having the wherewithal to push the button leads to a scene packed with the kind of imagery—action and emotion—that an English major could only dream of conveying.

But therein lies the rub. These most celebrated of images are each a bundle of choices—even if the only conscious choice was to plant oneself in a particular spot on a particular day. Ultimately, as much as we like to talk about “capturing” a scene, photos are (and have always been) something we create.

In subsequent classes, I’ll teach my students about the exposure triangle—how they can use the combination of shutter speed, aperture, and sensitivity to control things like the perception of motion, how much of a scene is in focus, and how light or dark an image is. An expert photographer might adjust these settings manually. A beginner may choose to shoot on “fully automatic” mode, but that doesn’t make the choices go away. The decisions are just being delegated to the machine. (Even an old-school, disposable film camera without a single piece of electronics inside has had the exposure parameters predetermined at the time of manufacture, and those parameters will still shape the appearance of the developed photo.)

One might try to imagine a world where there are no such mechanical considerations to be made—where what you see is simply what you get. Besides the fact that even our own eyes don’t work that way, that notion quickly fades once my students and I begin to discuss composition—how photographers (deliberately or not) make use of techniques like the rule of thirds and leading lines that subtly tell viewers what to look at, how, and when. We’ll talk about how the location of the camera—things like bird’s eye and worm’s eye view—can change our perception of a person’s size or power.

I love photography, so I could go on and on. (Don’t my students know it!) The point is that even when done in the service of journalism, a pursuit for which the central conceit is the primacy of truth, photography is at its heart artificial.

Creative Visions

We haven’t even begun to talk about portraiture. From the first moment some cave dweller smudged some dirt on a wall, we humans have been creating representations of the world around us. Walk through any art museum, and you’ll see paintings created to showcase this or that aspect of a person or a scene through the choice of medium, pigments, brushstrokes, and so forth, not to mention the complete control the artist has over what we see, what we don’t see, how large to make a person’s hip, etc., etc.

Portrait photography is really not so different. Whether in a studio or at a party, the photographer’s job is to capture the essence of one or more individuals and/or the scene around them. A wedding portrait may have the couple positioned just so with the lighting just right to portray their everlasting love for one another. A photo of a bunch of kids saying, “Cheese!” at a birthday party perhaps wishes to capture how gosh darn fun it is to be 7!

So is it really such a crime to use an algorithm to blend five images together so that everyone is happily looking at the camera if the whole point is to capture not just what the moment looks like but what it feels like—or even what we may want it to feel like?

Fowler’s column lingers on the opportunities for abuse, if not of this technology then its successors. But then again, as Jay Peters at The Verge points out, we’ve been manipulating photos almost as long as there have been photos to manipulate. (AI makes it a lot easier now, but the risk has always been there.)

What is it, then, that makes “Best Take” feel so different to me? Perhaps it’s the implied act of dissemblance—this sense that rather than just passing an image through a filter, we’re being given permission to pretend that a scene that we constructed from discrete parts (much like a LEGO model) equates to what we actually remember seeing in the viewfinder. What may be entirely natural to do with paint on canvas seems, in the end, subversive in a photograph.

Or, as Matt Bai puts it in his own Washington Post column, “It matters how we portray truth.” Writing about the entire range of “fun” new AI features users have access to, Bai goes on to point out the impracticality of trying to stop this technology. However, he adds, ominously:

But to make the explicit selling point of that phone the notion that imperfect truths don’t need to exist anymore—that what’s real is both fungible and subjective—strikes me as reckless. It romanticizes the most destabilizing trend in society and invites us all to revel in it.

Any neuroscientist will tell you that memory is malleable—even fallible. If our own perception of past reality can—and likely will—change over time, then there’s a part of me (albeit small) that wonders whether achieving a level of memorial perfection in photography even matters. Physicists, meanwhile, will tell you about the “observer effect”—the notion that the very act of observing a system will inherently change it. Bai’s larger political point notwithstanding, this poses a far deeper question: is it even possible to capture an “accurate” moment in time?

Maybe, as new tech often has the potential to do, what the Pixel 8 really inspires within me and the writers I’ve cited above is just good old fear of change—combined, in my own case, with my household’s disappointment over the Phillies’ loss in Game 7 of the NLCS.

I’m still undecided (about “Best Take,” not the unfairness of the baseball universe). But it is at least comforting to remember that, in a way, artificial intelligence in visual media is nothing new. Portraiture and photography have always been rooted in artifice. For better or for worse, we creators just have a lot more tools at our disposal now.