FTI Newsletter, Issue 101

Announcing: 2018 Summer Teaching Fellows Program
April 18, 2018
FTI Newsletter, Issue 102
April 29, 2018

Generative algorithms, faceprints, and putting words into other people’s mouths.

It’s been a while since our last edition of the Future Today Institute newsletter. Every two weeks, we take a deep dive into an emerging tech trend, exploring what it is, why it matters, key players, plausible scenarios for the future, and what action you and your organization should take.

Some weeks, the trend may not seem directly applicable to your work, however we encourage you to spend some time with it. To understand the future of one technology, you must consider the past, present and future of all things. Otherwise, you’re essentially looking at the world through a pinhole. For that reason, we hope you’ll allow yourself to let your mind wander productively.

“Stay woke, bitches.”

Don’t believe everything you see and hear in an internet video. This week, BuzzFeed and Jordan Peele teamed up to create a public service video that puts Peele’s words into President Obama’s mouth. It’s a 72-second digitally-manipulated video, in which Obama (who is actually Peele) says, “We’re entering an era in which our enemies can make it look like anyone is saying anything at any point in time, even if they would never say those things.” For example: “Ben Carson is in the Sunken Place” or “President Trump is total and complete dipshit.”

“Stay woke, bitches,” he says at the end.


Emerging Tech Trend:
Generative Algorithms For Voice, Sound and Video

Key Insight: Copious amounts of video, photo, and voice data are used to train machine learning algorithms. Over time, neural networks are capable of completing narrow tasks, like matching Jordan Peele’s voice to President Obama’s face.

Examples: In 2016, Adobe premiered a new tool called Voco, which it promised was a “Photoshop for voice.” The Face2Face algorithm allowed anyone to manipulate the face of anyone in a pre-recorded video. In 2017, researchers at the University of Washington developed a model that convincingly showed President Barack Obama giving a speech—that he never actually gave in real life. And late last year, there was a sudden rise of deepfakes—an evolution of these techniques that allowed anyone to face swap celebrities, coworkers, neighbors and family members into porn movies. Watch FTI Founder Amy Webb showing this technology in action last October at the 2017 ONA conference.

Near-Futures Scenarios:

  • Optimistic: Personalized Brand Voices. The new face of your company could very well be a synthetic voice and face. SF-based startup Voicery is one of the companies using machine learning and generative algorithms to create unique, synthetic voices. Think back to advertisements and commercials featuring Kate Moss for Calvin Klein, or Jennifer Gardiner for Neutrogena. Companies intentionally select people who embody their values and their promise to the consumer. Our modeling at FTI shows that within the next five years, 50% of the interactions we have with machines will be with our voices. This means that companies will need to develop a brand voice — and they’ll also have the opportunity to create a unique one, potentially for each consumer, using a recipe of different tones, cadences and inflections modulated in real time. Each of us could, theoretically, see and hear a slightly different version of Kate Moss, based on our digital behavior.
  • PessimisticAudio Fraud. Researchers at MIT’s CSAIL have been studying how to use machine learning and generative algorithms to automatically produce sound — not just human voices, but all of the ambient sounds that tend to follow us. For example, what sound is generated when a wooden drumstick taps a couch? A pile of leaves? A glass windowpane? The focus of this research is to help systems understand how objects interact with each other in the physical realm. But here’s an unforeseen consequence: those voices can also be hacked and repurposed. What if your company becomes the “voice” for a competitor? Or worse?
  • Catastrophic: Misinformation On Steroids. If the average person saw a video convincingly matching a generated voice with a generated face, would they be able to tell? Especially if that video was being rapidly passed around on Facebook and Twitter? We could be entering an era of hyper-realistic fake news videos, which could escalate political tensions and lead to catastrophic outcomes.

Action Meter: Organizations should spend some time thinking seriously about this trend and how it might mature. We advise you to think about generative algorithms across all of your business units. For those working in government and journalism, we urge you to consider how these tools could be used to spread misinformation.

 

Watchlist: MIT-CSAIL; Amazon; Arria NLG; Narrative Science; Expect Labs; Automated Insights; Department of Computing Science, University of Aberdeen; School of Science and Engineering, University of Dundee; Research Center on Information Technologies (CiTIUS), University of Santiago de Compostela, Spain; School of Informatics University of Edinburgh; Voicery; Lyrebird; IBM Watson; University of Washington; Apple; Microsoft; Google; Descript; WaveNet; Adobe.


Around the Institute

Summer Teaching Fellows: Applications are now available
Our 2018 Teaching Fellows Program is now accepting applications! This August, FTI will host a special Teacher Training Fellowship for teachers who want to incorporate futures forecasting into their curriculum. The program lasts three intensive days, and we will teach fellows how to incorporate the tools of futurists into their existing coursework. Apply here.

Futurist Amy Webb has some concerns about the future
Ahead of her talk at the Charles H. Wright Museum of African American History in Detroit, Crain’s spoke with Amy Webb about data privacy, local journalism and what we can do right now to make an optimistic version of the future come true. Read more…

FTI on TWiT
Mark Zuckerberg comes out of his Congressional testimony unscathed. China will dominate AI in the coming decade. HomePods are not selling like HotCakes. Apple leaks leakers leaking leaks. Waymo wants to test truly driverless cars in California. Watch or listen to “This Week in Tech” with host Leo Laporte, Lindsey Turrentine, Jason Hiner and Amy Webb. Watch/ listen…

Space Oddities: We Need a Plan to Stop Polluting Space Before It’s Too Late
We must learn to be better stewards of our own planet—and commit to very long-term thinking—before we try to colonize any others. Read more...

Get our tech trend updates

Comments are closed.