What could happen if our trust in video collapses?
Put another way, how could we respond when video can be altered as easily as images can be Photoshopped?
Let me explain. We are approaching that state now. Humans love video, both consuming and producing it. That video habit keeps growing, as a casual glance at Netflix, YouTube, or videoconferencing will show, and seems unlike to stall. At the same time, new tools are coming out, along with AI applications, which are giving us the power to remix video content in powerful ways. I’m not talking about FinalCutPro, which has let us edit footage and remix multimedia for years. I’m referring instead to the ability to alter preexisting video in ways that convince audiences of their veracity.
This has taken some time, as video files are larger and more challenging than images by an order of magnitude. Digital innovation just keeps racing ahead, and is approaching a crucial inflection point.
A few months ago RadioLab dove into this emerging trend. They showed multiple ways to alter video which might convince some audiences now. Check out the example they created: not that great, but could fool some folks. Or look at this work from 2016, nearly two years ago:
Two years is a long time in the digital world. After all, those tools are improving, and could well become more powerful in the short and medium term futures. We’re hitting the uncanny valley, and can scale the other side. Indeed, we’re starting to see early signs of this with people transposing faces (so far) onto those of porn actresses (and actors, I presume). “Deepfakes” is the keyword.
What about audio, though? Could video images become the prey of pranksters, but the human voice remain trustworthy? Probably not. Adobe is working on a tool currently dubbed VoCo (Voice Conversion, or Photoshopping Voiceovers) popularly dubbed “Photoshop for audio”.
It hasn’t appeared yet, but it seems likely that either Adobe will bring it to market, or someone else will offer their version, or both.
We can dive into these tools and approaches on a technical basis, but right now I’m curious about what their impact could be. What might follow the advent of, well, fake video, once it passes technical and popular tests? We can draw on the history of human responses to fakes for inspiration, as well as look to recent techno-cultural developments. After all, media provenance questions have been with us for millenia. (Consult Errol Morris on photography or Orson Wells on fakes, for two quick and terrific sources)
The battle between authority and user literacy expands One classic response to a credibility crisis is a turn towards authorities. If any video we might come across could be compromised or fabricated out of whole cloth, perhaps some skilled people can either determine accuracy to a satisfactory level. At the same time, certain sources could produce videos with a guarantee of their fidelity, setting up such certainly through, say, a transparent production process, or by presenting sufficient security to protect their work from alteration. This would offer the average user the option that leading newspapers presented a fake news concerned audience in 2017: trust in a handful of legitimate video outlets.
EDITED TO ADD: one way of establishing authority could be a new form of digital watermarking. Bethany Bovard had some ideas:
Do you think some form of watermarking is possible?
— Bryan Alexander (@BryanAlexander) January 29, 2018
On the other hand, another response to fake anxiety is to empower users/readers/consumers/patrons. That’s where literacies come in, from print to media, information to digital. This is one function of schools, and also of public pedagogy. Will we develop curricula to teach students how to detect fake videos themselves?
Will one of these approaches win out, or will they compete, or work together? (Last year I wrote about this authority<->literacy dynamic as a continuum.)
Keeping the faith Some may choose to maintain a trust in video fidelity, no matter what they hear (and see) about fake video (and audio). Millions of people now fall for faked photos, as a casual Snopes or Facebook trawl reveals. For an analogy, think how many people were appalled at what Snowden revealed about US government surveillance…. and then want right on with their lives without changing an iota of their digital behaviors.
Underground The opposite could happen. Instead of popular acceptance of a fallen video world, we could see mass revulsion. (For an analogy, consider the backlash to Google Glass, or the orchestrated marginalization of peer-to-peer file sharing.) That could drive fake video to the margins, even underground. In such a future we’d see sporadic stories of “photoshopped videos”, but they would be rare, and broadly hated.
Schism by video In some countries we see some people happily inhabiting media echo chambers, satisfied in being excluded from content and voices they abhor for various reasons. In the United States we see this reaching a fever pitch with president Trump’s denunciation of anything he dislikes as fake news, and some of his supporters apparently following suit. We’ve seen a similar pattern in the Philippines. If this is a useful antecedent, imagine a future with societies divided into competing ideologies or schools, each with their own fake-riddled video ecosystem. Partisans would scrutinize the opposition’s video objects strenuously, but relax when it comes to their side’s documents.
Open hostilities Today’s troll armies would seem likely to take advantage of fake videos. They already Photoshop images and create fake social media text content. Given the full power of video without constraints, imagine the epic rapes, tortures, deaths, etc. that these people could unleash.
Legal battles As both the Radiolab episode and the Wired articles linked above note, “Photoshop for video” means some interesting legal challenges. How much ownership does one have over one’s face or voice, for example? Remix has always been a tricky thing in copyright, and should become even more so. Will “revenge porn” laws expand?
A new career path People will make such tools and products. They could well do this as a business, or as part of a nonprofit. In short, fake audio and video could become a career path. Professionals (and amateurs) might require backgrounds in coding, multimedia editing, communication, and storytelling. Will schools of all sorts make avoiding such careers a part of their ethical establishment, or will some acquiesce to their students’ progress into the dark side?
Play and storytelling The flipside of fake video terror is playfulness. For example, when we look at today’s concern about disturbing YouTube videos aimed at children and clicks, we shouldn’t ignore the manic surrealism displayed in those many, many recombinations of characters, objects, and settings. Surrealism is a style that hasn’t gone away in the century since its inception (another WWI output, by the way). It appears in popular movies, avant garde art, advertising, and games. If our taste for it persists, we should expect people to “Photoshop videos” into new realities, like ours, just slightly twisted.
History can be a guide to this. Think, for example, of Max Ernst’ Une semaine de bonté (1934), with its wild, disturbing, and funny transpositions of heads, dragons, and sculptures across contemporary popular graphics.
…and that’s a start. How do you see the era of fake audio-visuals playing out?