Is there a hard line between real and fake? I'm a real person. My friend Jenny is a real person, too, even though a part of her elbow is held together with some screws and pieces of metal. My friend Rachel is a real person, too, even though a battery-powered device helps her heart beat at the right tempo. To continue into the (for now) fictional: Inspector Gadget (from the 1980s cartoon series) had innumerable machine parts inserted into his human body in order to give him special abilities.
When this becomes possible, will Inspector Gadget be considered a real person? In the D.C. Universe, there's a character named Metallo, who consists of a human brain in a robotic body. Would Metallo be a real person? In the Body of Theseus, how much of a person could be replaced by non-original parts before that person wasn't considered human anymore?
A little over a year ago, there was some controversy over the cameras in Samsung phones "faking" high-resolution photos of the moon. It's explained very well in this article by James Vincent and Jon Porter, but the basic idea is that S23 Ultra phones appear to sharpen blurry images by using artificial intelligence to enhance photos of the moon. This is similar to what photocopy machines have been doing for years by upscaling images, but Samsung isn't just upscaling; it's providing the images with new details that weren't actually captured by the camera lenses themselves. Is the image on the right fake?
The Verge article I linked to earlier maintains that "the concept of 'fakeness' is a spectrum rather than a binary." At one end is the sex tape of Gal Godot with her stepbrother mentioned in this New Yorker article about deepfakes (paywall? Try https://12ft.io/), and at the other end is...I'm not sure. Can you even take a photograph that isn't, in some way, "fake"? A couple of weeks ago, I was seeing really impressive pictures of the aurora borealis over the United States on social media; it was visible this far south because of intense solar flares. I went outside, and even drove a few miles away from town to get away from light pollution to see the sight for myself. It was disappointing, to say the least. Apparently the great photos that my friends were taking were not representative of what they were actually seeing that night with their eyes. Were their pictures fake?
There's an episode of the podcast Radiolab that tells the story of Wilson Bentley, a teenager in the 1880s who took the first photographs of snowflakes. He did so in an unheated shed, using a camera attached to a microscope. The photos were beautiful; and they were loved by both scientists and laypeople. According to the podcast, however, it was soon revealed that Bentley had altered his images. He used a knife to "scrape the negative around the snow crystal" to create a dark background and make the image of the snowflake stand out more, with sharper edges. Bentley's response to his critics was, "A true scientist wishes, above all, to have his photographs as true to nature as possible. And if retouching will help in this respect, then it is fully justified." It turns out that many early photographers edited their images for various reasons, one of which was to make up for the fact that cameras weren't all that great when they were first invented, and they wanted photographs to look more like real life looked. Where does this fall on the spectrum of fakeness?
Did the Beatles really release a new single last November? Where does the use of posthumous recordings fit, especially ones that use artificial intelligence? There aren't any recordings that are "real," any more than there are real photos. Whose "real" live listening experience would one be trying to replicate when recording, say, an orchestra, anyway - the conductor's? An audience member in the fourth row center? Every person in every space would have a different live experience, so what is a "real" recording? When The Beatles Anthology (1995) was being made, "Now and Then" was being considered as a new release along with "Free as a Bird" and "Real Love." Paul McCartney, George Harrison, and Ringo Starr even went as far as to record some new parts before it was abandoned, due to the quality of Lennon's original tape. It just wasn't possible to isolate his voice and piano parts and get rid of background noise in the ‘90s, and this is where the artificial intelligence comes in: we can do it now. To be clear, AI didn't write the song, although it might have been able to. AI didn't recreate John Lennon's voice or George Harrison's guitar playing, although it probably could have done that, too. All it did was clean up the recording so that Lennon’s vocals and piano parts were useable. So...is it fake?
To return to Vincent and Porter once more, they claim that "Ultimately, photography is changing, and our understanding of what constitutes a 'real photo' will change with it." Is this true? Is it true about music, too? And does it even matter? Daniel Immerwahr (The New Yorker, again) argues that despite general panic surrounding deepfakes, they haven't made an actual difference in the outcome of any election. "Social verification" helps us decide what's real and what isn't. One person might be fooled, but "We are, collectively, good at sussing out fakes..." Immerwahr concludes that figuring out what is fake and what is real is not the most important task at hand. Rather, what people are choosing to fake is revealing something about them. "The problem with fakes isn’t the truth they hide," he says, "it’s the truth they reveal."
Comments