BY KIM BELLARD
The Tom Cruise TikTok Deepfakes last spring didn’t encourage me to write about deepfakes, not even when Justin Bieber was so in love with her that he was challenged the deepfake to fight. When 60 minutes the subject covered Last night I figured I’d best get there before I missed this particular wave.
We are already living in an era of unprecedented misinformation / disinformation, as we have seen repeatedly with COVID-19 (ex.
ICYMI, here it is 60 minutes Story:
The trick behind deepfakes is a type of deep learning called “Generative opposing network“(GAN), which basically means that neural networks compete for which of the most realistic media (e.g. audio or video) can produce. You can try to recreate a real person or create completely fictional people. The more you iterate, the more realistic the output becomes.
Audio deepfake technology is already widespread, and already pretty good. The software takes a sample of a person’s voice and “learns” how that person speaks. Type in a sentence and the software will produce audio that sounds like the real person.
The technology was already used to outsmart a manager about sending money to an illegal bank account by deepfaking his boss’s voice. “The software was able to imitate the voice, and not just the voice: the tonality, the punctuation marks, the German accent,” said a company spokesman told The Washington Post.
You have to assume that Siri or Alexa would also fall for such deepfaked voices.
Audio deepfakes are scary enough, but video takes it to another level. As the saying goes: seeing is believing. An expert in cybercrime told The Wall Street Journal: “Imagine a video call with [a CEO’s] Voice, the facial expressions you are familiar with. Then you would have no doubts at all. ”
As is so often the case, the porn industry is an early adopter of the new technology. Last month MIT technology review reported on a website that allows someone to upload a picture of a face and see that face transformed into an adult video. The effects on innocent victims are terrifying.
This particular site (the Technology review now says is no longer available) wasn’t the first such porn site to use the technology, probably hasn’t had the most realistic deepfakes, and won’t be the last. Unfortunately, however, deepfake porn is nowhere near the biggest problem we’re likely to have with technology.
We’re going to see mainstream actors in films they never made. We will see dead Actors in new films. We’re going to see Deepfaked executives saying all sorts of ridiculous things (Mark Zuckerberg may already be a deepfake). We’ll see politicians say things that make their opponents look good.
Martin Ford, enroll Market observation, warns:
A sufficiently believable deepfake could literally shift the arc of history – and the resources to create such inventions could soon be in the hands of political activists, foreign governments, or just mischievous teenagers.
Hany Farid, a Cal Berkeley professor, told NPR: “Now you have the perfect storm. I can create this content easily, cheaply and quickly, I can deliver it to the world en masse, and I have a very willing and eager audience that will reinforce this for me. “
Likewise, technology consultants Nina Schick, who wrote a book about deepfakes, told 60 minutes: “The fact that AI can now be used to create fake pictures and videos that look hyper-realistic. I thought this was a game changer from a disinformation perspective. “
Imagine what the COVID misinformation crew did with a deepfake Dr. Fauci could hire.
He was in many ways the face of modern medicine and science during the pandemic. There have been countless hours of video / audio from him over the past 18 months. Usually he was right and sometimes wrong, but he did his best to follow science. COVID-19 skeptics / deniers constantly analyze his words in search of inconsistencies, times when he was wrong, after every opportunity to question his expertise.
With deepfakes, he could tell people not to bother with masks or even vaccines. His deepfake could advertise unproven and even unsafe cures and denounce the FDA, CDC, even President Biden. Hell, they could get President Biden to call Dr. Attacking Fauci and praising Donald Trump (conversely, a deepfake Trump could of course push for vaccine mandates).
We are now striving to find the best health information about COVID and anything else that worries us about our health. We look for credible sources, seek the opinions of reputable people, and use this information to make our health decisions. But as Frau Schick said further 60 minutes‘Deepfakes will “require all of us to figure out how to maneuver in a world where seeing isn’t always believing.”
That will not be easy.
We’re just beginning to see how deepfakes can affect healthcare. In one recently nature items, Chen, et. alias warned:
… the proliferation of deepfakes is a blind spot in healthcare; Current measures to protect patient privacy, authentication and security are inadequate. For example, deepfake generation algorithms can also be used to potentially impersonate patients and exploit PHI to falsely bill health insurers that rely on imaging data to approve insurance claims46 and manipulate images sent from the hospital to an insurance provider to trigger a reimbursement request for a more expensive procedure.
The authors believe that synthetic data has a role in healthcare, but say: “There is an urgent need to develop and refine a regulatory framework using synthetic data and monitoring its impact on society.”
It is generally like that. The technology used to detect deepfakes is improving, but so is the technology to create them. It’s an arms race, like everything in cybersecurity. As Ms. Schick further emphasized 60 minutes, “The technology itself is neutral.” How it is used is not.
But she also believes: “It is without a doubt one of the most important revolutions in the future of human communication and perception. I would say it’s analogous to the birth of the internet. “
I’m not sure I would go that far.
Revised audio / video was with us pretty much all the time we had audio / video; Deepfake technology takes it to a new and more compelling level. We still haven’t figured out how to use the internet responsibly, and when they stop doing something, deepfakes remind us that we’d better be doing this soon.
Kim is a former e-marketing manager on a major blues plan, editor of the late & defendant Tincture.io, and now a regular THCB contributor.
Thank You For Reading!