child holding a photograph of someone else's lips in front of their mouth
PHOTO: Christian Gertenbach

“A lie can travel halfway around the world before the truth can get its boots on.”

The quote above is often attributed to Mark Twain. Unfortunately, there's little evidence he said it (read the full background story, including the possibility Jonathan Swift was the source, rather than Mark Twain). Which gives us a good place to start a conversation about fake content.

Innovative, Yes. Progress? Who Knows

As we push the frontiers of technology, the law of unintended consequences frequently rears its head in unexpected ways. Sometimes, the unintended consequences of technology innovation takes the form of unexpected benefits — such as the network effects of everyone having a mobile device. At other times, it manifests as unexpected drawbacks — like our Wi-Fi connection to HBOGo shifting to a lethargic crawl as a result of everyone tuning into the finale of Game of Thrones at the same time.

But the worst unintended consequence is an effect contrary to what was originally intended, when an innovation makes a problem worse.

For example, being able to connect with anyone, anytime, anywhere and without filters hasn’t exactly liberated our better natures. As I mentioned in a post last year, the lofty ideas of peace, love and understanding promulgated by Facebook, Twitter and YouTube/Google have clearly collided with three unanticipated forces:

  1. Web business models based on ever-more niched and segmented click-based advertising.
  2. A regulatory infrastructure (Section 230 of the Communications Decency Act of 1996) that says, "No provider or user of an interactive computer service shall be treated as the publisher or speaker of any information provided by another information content provider."
  3. Nefarious players on both the left and right, augmented by nation-state resources, determined to exploit No. 1 and No. 2.

We haven’t exactly enjoyed the unintended consequences of the intersection of Fake Content 1.0 with these unfiltered and advertising-driven social platforms. We’ve watched as players as varied as nation-states like Russia and China and online cranks with a grudge and a laptop have manipulated these infrastructures to sow disruption.

Related Article: Crazies at the Social Media Barricades: Unintended Consequences and Next Steps

AI Technology Is Surpassing Our Ability to Respond

Just a few weeks ago, a video purported to show House Speaker Nancy Pelosi slurring her words got millions of views in the blink of an eye. And despite all sorts of hand-wringing, most efforts to stop it likely only made things worse.

That’s because a lie can travel halfway around the world before the truth even gets its boots on. The fact that all of this was caused by a pretty ham-handed fake video — creative editing plus slowing down segments by 25% — ought to scare us all. We haven’t exactly done a good job of sorting authentic from fake in the Fake Content 1.0 era, whether it be videos like the Pelosi video or the photoshopping of celebrity heads onto porn star bodies.

All this is just the tip of the fake content iceberg as AI now meets video. A 2016 article in Wired, "How Photos Fuel the Spread of Fake News," almost seems quaint now.

The technology itself is fascinating. But the implications are complex. And unfortunately, our ability to create guardrails to control the technology will likely lag behind the pace of technology itself for the foreseeable future.

Related Article: Is it Time for Your Organization to Form an AI Ethics Committee?

Speeding Towards an Uncertain Future

For example, take a look at Supasorn Suwajanakorn's TED video from last year. Suwajanakorn is a computer scientist. In his talk he shows how, as a grad student, he used AI and 3-D modeling to create photorealistic fake videos of people synced to audio, focusing on the positive and creative uses of the technology.

“Here's Richard Feynman, who in addition to being a Nobel Prize winner in physics was also known as a legendary teacher. Wouldn't it be great if we could bring him back to give his lectures and inspire millions of kids, perhaps not just in English but in any language? Or if you could ask our grandparents for advice and hear those comforting words even if they're no longer with us? Or maybe using this tool, book authors, alive or not, could read aloud all of their books for anyone interested. The creative possibilities here are endless, and to me, that's very exciting.”

Or this approach — Deep Video Portraits — SIGGRAPH2018 — presented at last year’s SIGGRAPH conference.

“We present a novel approach that enables photo-realistic re-animation of portrait videos using only an input video. In contrast to existing approaches that are restricted to manipulations of facial expressions only, we are the first to transfer the full 3-D head position, head rotation, face expression, eye gaze and eye blinking from a source actor to a portrait video of a target actor. The core of our approach is a generative neural network with a novel space-time architecture.”

Yet another innovation comes from researchers at Carnegie Mellon University, who have “created a method that automatically transforms the content of one video into the style of another, making it possible to transfer the facial expressions of one person into another. For instance, Barack Obama’s style can be transformed into Donald Trump.”  

Oh good.

Only three weeks ago, the Samsung AI Center at Moscow and the Skolkovo Institute of Science and Technology published a paper under the super sexy title, "Few-Shot Adversarial Learning of Realistic Neural Talking Head Models." The researchers noted they’d overcome the need for the huge amount of raw photo images usually required as AI inputs to create a video. The result was a pretty convincing video created from as little as a single image and that for “perfect realism” the model was trained on only 32 images.

Lest we take some comfort that at least all of these technology innovations are still largely only within the reach of responsible super technical peeps, like any cloud technology, these technologies are also rapidly being consumerized and increasingly made available to anyone, largely driven by — you guessed it — porn.

So. What do we do about this? Most of the innovators offer reassurance like this from Supasorn Suwajanakorn, which I sure hope winds up being well-founded:

“Our goal was to build an accurate model of a person, not to misrepresent them and I'm part of an ongoing effort at AI Foundation, which uses a combination of machine learning and human moderators to detect fake images and videos, fighting against my own work. One of the tools we plan to release is called Reality Defender, which is a web-browser plug-in that can flag potentially fake content automatically, right in the browser. But I'm excited and hopeful, because if we use it right and carefully this tool can allow any individual's positive impact on the world to be massively scaled and really help shape our future the way we want it to be.”

Frankly, I don’t know. I do know that the abuses and excesses of rapidly evolving technology tend to outpace our ability to control it.