DeepFakes

What do we do when we can’t trust our eyes?

The Viral Nancy Pelosi Slurred Speech Video

The line between what is real and what is fake is getting blurry. For all the talk about “Fake News,” the last two years, it is easier than ever for misinformation to spread across social media and the web.

On this Memorial Day weekend, some readings and reflection on DeepFakes and the uncertain future of “reality.”

Articles:

1. Faked Pelosi videos, slowed to make her appear drunk, spread across social media — Washington Post
2. Fake video makes Nancy Pelosi look drunk. Facebook won’t take it down. -Vox
3. Bill Hader’s Al Pacino impression with the help of deepfakes — Fast Company
4. AI Recreate Podcast Host Joe Rogan’s Voice To Say Anything- PCMag
5. Facebook Removed Over 2 Billion Fake Accounts, But The Problem Is Getting Worse — Buzzfeed News

Read widely. Read wisely.

Max

P.s. I’m including a mailbag section at the end for a few of you who’ve written back to recent editions.

subscribe here

Recommended Readings

1. Faked Pelosi videos, slowed to make her appear drunk, spread across social media
Washington Post (5 minutes)

SUMMARY
Distorted videos of House Speaker Nancy Pelosi (D-Calif.), altered to make her sound as if she’s drunkenly slurring her words, are spreading rapidly across social media, highlighting how political disinformation that clouds public understanding can now grow at the speed of the Web.

SELECTION
One version, posted by the conservative Facebook page Politics WatchDog, had been viewed more than 2 million times by Thursday night, been shared more than 45,000 times, and garnered 23,000 comments with users calling her “drunk” and “a babbling mess.”…

Rudolph W. Giuliani, Trump’s personal attorney, tweeted a link to the altered video Thursday night with the note, “What is wrong with Nancy Pelosi? Her speech pattern is bizarre.” The tweet has since been deleted.

2. A fake viral video makes Nancy Pelosi look drunk. Facebook won’t take it down.
Vox (5 minutes)

SUMMARY
Facebook has taken all kinds of heat for hosting false political information.And rightfully so. Now it invests millions into scrubbing the content on it’s site. But just because something isn’t true doesn’t mean they will take it down.

SELECTION

Facebook’s third-party fact checkers reviewed the video on Thursday eveningand rated it as false, and now the company is reducing its distribution and showing additional context alongside it in the form of a related article in news feeds where it appears. But Facebook’s community standards don’t require that information posted on the platform is true. A spokesperson for Facebook declined to comment on the record for this story.

3. Bill Hader’s Al Pacino impression gets even more real (and creepy) with the help of deepfakes
Fast Company (4 minutes)

4. This AI Can Recreate Podcast Host Joe Rogan’s Voice To Say Anything
PCMag (4 minutes)

SUMMARY
The day when technology can recreate your voice to say anything — no matter how insane — is pretty much here. A group of engineers has created an AI program that can convincingly mimic the voice of podcast host Joe Rogan.

SELECTION
You can hear the results in the video in the link. The voice is not only identical to Rogan, but it can also produce a natural-sounding rhythm to almost every word said.

However, the voice is artificial. The engineers at the AI company Dessa can get their fake Joe Rogan to say anything — all they have to do is type in the text. In this case, they made the AI-powered voice talk about sponsoring a hockey team full of chimpanzees and also tout the benefits of being a robot.

The video will probably both amuse and disturb you. And that’s the point.

5. Facebook Removed Over 2 Billion Fake Accounts, But The Problem Is Getting Worse

BuzzFeed News (8 minutes)

SUMMARY
Facebook announced this week it removed 2.19 billion fake accounts between January and March, its biggest-ever takedown of fakes in a single quarter. Company CEO Mark Zuckerberg confidently told media that “we’re taking down more fake accounts than ever.”

SELECTION

What he didn’t say is that there are also more active fake accounts on Facebook than there were six months ago — more than ever before. Facebook now says 5% of active accounts are fake, up from its previous estimate of 3% to 4%.

The sweep of more than 2 billion fake accounts, while impressive, mostly removed profiles at the point of creation. As Zuckerberg said, “they were never considered active in our systems and we don’t count them in any of our overall community metrics.”

So based on Facebook’s own key metric for measuring fake accounts, the problem is actually getting worse. People who investigate fake accounts on the platform also told BuzzFeed News that the company still fails to remove obvious fakes — even during critical periods, such as the ongoing European Parliament election.

Postscript

There is a scene in Forrest Gump when Forrest goes to the White House and meets JFK. The filmmakers edited in Tom Hank’s face to appear in archival footage of the president. At the time, it looked completely real — like he was really meeting Kennedy. I remember seeing it in the theater for the first time and being blown away. This changes everything, I thought. If they can make this look believable, they can do anything.

Of course, back then, the technology to alter the film was new and extraordinarily expensive. Not today. With tools you can download this afternoon, you can create videos of people saying things they never said and the quality would be good enough to fool your friends.

These videos are called deepfakes. They use deep-learning artificial intelligence to create believable fakes of people saying whatever the creators want them to say. The NancyPelosi video isn’t actually a deepfake, it’s a crude alteration that just slowed the video down. As we head into 2020, I expect we’ll see dozens more and the newscycle will be filled with fake videos of candidates followed by their statements that the videos weren’t real.

Now, to be balanced, the rise of DeepFakes aren’t necessarily cause for panic. People have been altering and manipulating photos for years and we’ve learned to live with that, deciphering the fake from the real. Video is just the next step. On the other hand, I’m not sure living with photo alterations has been harmless. As a father of daughters, I don’t like that they will grow up looking at retouched photos of other women, giving them distorted ideas of the female body.

The optimistic take on The Information Age has always been that the web opens up an unlimited opportunity for accessing knowledge. The pessimistic take is that it also opens up an unlimited opportunity for spreading misinformation. Russian interference in the 2016 election is only the most well-publicized example. There are others.

• Axios reports that “A Pew survey last year found that two-thirds of tweeted links to popular websites came from non-human users (bots or other automated accounts).”
• Another study found that more than half (half!) of all web traffic was from bots, not humans.
• Anti-vaccine content is proliferating online

Meanwhile, Facebook is beginning to censor individuals they consider dangerousfor spreading misinformation, recently banning conspiracy theorists like Alex Jones, among others. In other words, without any meaningful government oversight, Mark Zuckerberg has a power that few have imagined in history: the power to determine what is helpful and hurtful to the public, even more, to determine what is true and what is false. That kind of power is what has led Zuckerberg’s co-founder Chris Hughes to argue it is time to break up Facebook.

The AI Foundation has created a browser plugin called Reality Defender “built to run alongside digital experiences (such as browsing the web) to detect potentially fake media. Similar to virus protection, it scans every image, video, and other media that a user encounters for known fakes, allows reporting of suspected fakes, and runs new media through various AI-driven analysis techniques to detect signs of alteration or artificial generation.”

Just think of that. We live in an era where truth is so much in question that we feel compelled to defend reality.

In a way, The Weekend Reader, is my low-tech version of defending reality. Instead of sharing the hottest link and most viral tweet, I try to take my time, only sending once a week. And instead of relying on one source, I share multiple articles, from a variety of sources. The hope is that the variety will give a fuller picture of reality and help us all think a little better about what’s actually true. Someday soon there will probably be an AI that can do it better than me. But until then, thanks for reading.

Read widely. Read wisely. Don’t believe everything you see.
Max

MAILBAG

I thought I’d share a couple of the responses to last week’s issue. If you want to respond to this week’s issue, drop me a line! No promises that I’ll publish it or even reply, but it’s fun to keep the conversation going.

On Last Week’s Sugar Issue (LINK)

MAT writes: Two things… 1. Where does Shake Shack fit into this? And 2. So glad they know about Fit Fat people being healthy. Makes me feel fat strong.

• MAX: 1. Shake Shack is like Doritos and Ice Cream. So delicious I think it’s foolhardy to build a diet that excludes it. 2. I like the Fit Fat club too! #chartermember #fatstrong

***
JAMIE writes: The best book I’ve read [on this topic] has been The Obesity Code. You may enjoy it. Connects a lot of thoughts together and gives a solid plan. https://www.goodreads.com/book/show/24945404-the-obesity-code

• MAX: Thanks for the recommendation!

***
JEREMY writes: I love your selections. I read them every week. But as a physician, I had a problem with #2. Sorry, man. Just sayin’. I don’t know many — no, any — obese doctors. Sorry, but I just don’t. They’re probably out there. But I don’t know any.I also lived in Ethiopia for 4 years. Guess what? I didn’t see any fat patients in Ethiopia. None. The people there are skinny. They have malaria and TB and HIV. But in general, they don’t have strokes and heart attacks.

It’s hard for me to believe that obesity and patients being overweight is not a Western problem. And a problem of an excess of available cheap calories. Sorry, but I disagree with Mr. Hobbes wholeheartedly.

It is true that diets do not work. But it is not because being slimmer is not healthier. It is! It is because diets are extremely hard to abide by in our Western culture. I know because I weighed 20 pounds less when I lived in Ethiopia. In the last 4 years, I have done keto, paleo, you name it to remove the excess 20. But I can’t. You know why? Because just when things are going well, I eat a burger and onion rings, and a bag of chips, and it’s all downhill from there.

Do some obese people have normal blood pressure and normal cholesterol? Sure. Do some skinny people have high blood pressure and high cholesterol? Sure. But it’s an undebatable scientific fact that obesity in general leads to poor health outcomes, and the author of this piece cannot dispel that no matter how eloquent he is.

• MAX: I’m glad you wrote this. Good to have some back and forth on an idea as controversial as “you can be fat and be healthy.”