> It's a time when we need to rely on trusted news sources
That's the rub, isn't it? Who's a "trusted" new source? All the major players have shown biases, corruption, and manipulation. Lots of people therefore consider some fly-by-night facebook feed that says what the majors don't to be more "trusted" in response.
We shouldn't "rely" on single points of perceived trust, especially in this massive for-profit, for-power media industry. We need to be broadly informed from various angles to have a better chance of piecing together a reasonable sense of what's going on.
The problem lies in the motive. News organizations are (not surprisingly) entirely motivated by profit. Breaking even is not enough. Doing well is not enough. Improving profit drastically quarter over quarter is the goal.
To do that, you cannot simply "provide news" anymore. You need conflict. You need confirmation bias. You need to create a wild, screaming echo chamber that induces panic and fear. You need conflict, even if you have to fudge the details to get it. You need to focus on some things and ignore others to increase it.
It's no longer a partisan issue. I'm 40 years old. I remember when you had to really pay attention to catch bias in the news. Do this yourself: go back and look at newsreels from the 80s and 90s. It's almost surreal. They were still biased, but presented information in a factual manner and then tossed in some emotional stuff. Many of them even... showed both sides of the argument. It's weird to watch now.
A news network simply cannot run with this format in 2018. Fox News was the leader, they were using bias, fearmongering, and propaganda almost from the start. They enjoyed very high ratings because of it. Now the other side has caught up. Because they had to.
It's a mess, but one thing is for sure: we can't trust any of them anymore. Not 100%.
Go back even further and pretty much every news source was a propaganda machine. Most of the major "trusted" new sources today started primarily in yellow journalism and political attack ads. Pulitzer was not a particularly nice man. Neither was Hearst.
It was a brief modern period when news sources decided to take themselves seriously and rise up ethically and appear "unbiased" to the public. It made the landscape more consolidated and monolithic, but it also drove sources to present relatively whole stories with what limited bandwidth they had to the largest population possible.
That being said, at the same time, radio hosts quite willingly went into the base-pandering and echo chambers. Conservative Talk Radio didn't crop up in the 2000s. They arrived after the repeal of the Fairness Doctrine from the FCC in 1987.
TV and print news only later realized if you insert just enough bias to make their viewers feel comfortable and to differentiate from their peers, they could make a WHOLE lot more money. And with the creation of a lot more time to sell news via 24hr channels and the Internet, you had to room to add in fluff, opinion, and other material.
> News organizations are (not surprisingly) entirely motivated by profit.
That's an over-simplication.
Sure, Jeff Bezos probably wants to make a profit with his purchase of the Washington Post. But the profits there are probably not his main objective as they are too small to make much of a difference to a person of Bezo's wealth.
Bezo's goal is to have a mouthpiece to push his agendas, such as pushing Amazon and trashing Musk properties. So you say that the purpose in doing that is to make more money. But the problem with that line of thinking is that you could reduce almost any action to "profit". You could reach further back and go a level lower and say that making profit is motivated by reproductive fitness, and thus "News organizations are entirely motivated by sex". But at some point you have to let go of the circuitous route from action to final motivation and focus on the immediate drivers of an action.
Politics, money, sex, job security, prestige, etc all all motivations for news organizations, as well as many other organizations.
It’s a question of mindset, but often correlated with scale. A rich person is a successful car dealership owner who has a net worth of $50 million, drives a Lamborghini on the weekends, lives in an 10,000 square foot mansion, has a sweet private movie theater, and owns a 60 foot powerboat. His economic activity is focused on consumption. He wants to leave his kids enough money to be rich for life. I can’t give many examples because these people are soon forgotten, but Donald Trump was almost one of these guys until he got some new ideas late in life.
A wealthy person has a net worth of $100 billion, is driven around in a Mercedes S600, has a 120 foot sailboat, goes to private parties at Sundance, values their time above everything else, and is focused on his legacy and place in history. His economic activity is focused on investment. He wants to leave his kids with enough money, connections, and capabilities to become among the great American families, shaping policy and public life for centuries. Think Rockefeller, Carnegie, Gates.
As a political radio host I used to listen to once said (on air, paraphrased from a memory from the early 2000s): "You should doubt and check everything I say. My job is not to inform you or tell you the truth. My job is to keep you listening so advertisers will pay my employer to run their ads."
Turning 40 this year, so I remember the same media you do. As the Internet grew and options proliferated, I always wanted a particular kind of news that I haven't seen crop up yet. Just a feed. No commentary, no talking heads discussing what something might mean. Just a network which shows video of news events happening. If something is happening in the legislature, I want to watch the legislators doing it. I can get the interpretation and more information about what's going on somewhere else. But the bare "this is what happened" without the fluff just isn't something you can really get. And sure, there's still editorial influence in what they would choose to show and how its shown and other such things, but a countless number of times I've just wanted to see the thing that actually happened.
We have known for centuries that who says a thing cannot influence its truth. How a thing is expressed can't affect it either. An untreated schizophrenic hobo screaming into the street can be telling the truth, and an esteemed professor reporting on new findings in his field can be a liar. It has always been inadvisable to do anything other than analyze claims critically, and I suppose if deepfaking becomes widespread, that will have to be relied upon more than ever.
Instead of your hypothetical ideal online news feed, we get the opposite: headlines which now attempt to directly program the reader's attitude about the thesis of the article, because nobody reads the article anyways these days (in this information economy), just the headlines. Thus, the body of the article is increasingly useless as a persuasion tool, especially compared to the headline, so the result--the headline which tells you the thesis of the article and how to feel about it in the same breath--is only the logical conclusion to what social media is doing to our brains, and to society.
> I remember when you had to really pay attention to catch bias in the news.
I wonder about this. Part of the reason you had to pay attention was that there were far fewer information sources to check against. No one was reporting on Martin Luther King's affairs, and no one was reporting on the FBI's attempts to get MLK to commit suicide over them either.
That strikes me as some sort of backwards reasoning, people had to literally break into FBI offices and steal documents so that those things would come to be known. You can't really fault the news for not reporting on something they had, unless I'm mistaken, absolutely no way of knowing unless they were to, y'know, infiltrate the FBI.
The burglary was in 1971, and the MLK letter didn't get published until much later. These days the documents would get uploaded to Mega or sent to Wikileaks, and dozens of people would pore over them.
It's not even a question of partisan bias; it's a question of sensationalism. It doesn't begin and end with how news organizations cover events; it begins with what events they choose to cover.
Awesome interview, thanks for the link. The end of the interview where Chomsky describes the Internet as a double-edged sword and (correctly) suggests that it could turn into "a home-marketing service, and a way of marginalizing people even further" is particularly interesting.
> All the major players have shown biases, corruption, and manipulation.
Bunching those together feels disingenuous. Do the New York Times, Washington Post, WSJ, etc etc all have their own perspectives? Sure. That's bias, to an extent. But when have they been corrupt? When have they manipulated their reporting?
To put it simply: if a reputable outlet was caught modifying a video of Trump to put words in his mouth it would be a publication-ending scandal. The same is not the case for Infowars. I agree with you that everyone should have as diverse a media diet as they can, but suggesting that every company is as bad as each other is simply inaccurate.
The magnitude of incorrectness/malicousness in reporting is irrelevant to this: No single source will leave you with a broadly-informed understanding, even if by simple way of time/space constraints in their output.
Even the simplest of biases affects editorial decisions as to which content to prefer over others. Omissions of viewpoints or of counterpoint example stories means you don't get the big picture, even without some massive media scandal. People scream distrust at media channels for leaving out information they view as critical, and go elsewhere to find those details (for better or worse).
> To put it simply: if a reputable outlet was caught modifying a video of Trump to put words in his mouth it would be a publication-ending scandal.
Caught how? You could show people the original video, but the reputable outlet and its allies would be pushing the narrative that the so-called "original" video is a deep fake produced and pushed by conspiracy theorists. After all, our outlet would never lie, would they?
I suspect the scandal would be less publication-ending than you think.
>>But when have they been corrupt? When have they manipulated their reporting?
I think this is your bias showing if you believe the New York Times, Washington Post, and WSJ are not just a manipulative in their reporting as CNN, Fox News and MSNBC.
Do they outright lie, of course not, but they omit things, they choose not to run some stories that do not fit their narrative, they choose to "expose" the most extreme version of the story.
The Current Gun Rights debate highlights this nicely for all of them.
I think CNN was clearly corrupt in the last election cycle. That's pretty well documented (especially helping Hillary against Bernie). It wasn't just "manipulative in their reporting". It was out-and-out helping a campaign win.
Are the NYT, WP, and WSJ biased? Sure. Are they corrupt in the sense that CNN was? I haven't seen evidence of it.
If anything the Washington Post was worse than CNN. They ran 16 anti-Bernie stories in 16 hours [1]. Their outright complete support of Hillary was very obvious.
NY Times often used spurious polling and statistics to push Hillary. They messaged that Hillary was in the lead before primaries even began based on predicted super delegate numbers. They claimed with over 90% certainty that Hillary would win the election on election day, even when it was obvious to casual observers such as myself that that number was totally wrong. They were trying to manufacture consent.
> They messaged that Hillary was in the lead before primaries even began based on predicted super delegate numbers.
Which was true.
> They claimed with over 90% certainty that Hillary would win the election on election day, even when it was obvious to casual observers such as myself that that number was totally wrong.
Even election day doesn't necessarily prove that "wrong". 90% chance of Hillary winning means a 10% chance of Trump winning. If I draw a red ace of clubs out of a deck of cards, I don't go "that's statistically unlikely, fake news!"
It appears that 538's model was better, in hindsight, as he appears to have been correct to give more credence to the chances of a multi-state polling miss. It's not clear the NYT's model was necessarily wrong, through.
You are correct that all sources have bias. It's difficult for readers to check every article for this and gauge their credibility. So my startup, CivikOwl, built a Chrome extension that makes it easy to check the quality of citations in an article and scores it based on this and other signals.
If you're interested check it out: http://www.civikowl.com. Feedback welcome (If you want to check out beta mobile app hit me up: arjunm at civikowl dot com)
(There's also a whole business model behind this that helps news orgs move away from reliance on ads but that's a longer discussion).
Yes, the requirement isn't that a piece of media came from some digitally signed hardware, but the process of recording involved (idk) sampling from a real world source.
I don't know how you would do that in the absence of the real world source, since the recording event would no longer exist.
Here's what I'm worried about with fake video. If you think fake Russian bot news in US politics is a problem, consider the following.
Facebook is increasingly populated with a large number of new users from developing nations who are basically one or two generations removed from subsistence farming. The level of naivete and rumor spreading that you can see in Pakistani Facebook is alarming. Now combine nation-state funded agitprop campaigns with naive users and fake video. Read the following article and imagune how much worse it could be if a malicious organization with funding decided to further weaponize content on myanmar-oriented Facebook groups and pages.
Why should we assume that being one or two generations removed from subsistence farming implies that a user is naive? There are plenty of naive persons worldwide who grew up in middle class households in the West who are just as naive. I don't think the connection is that clean.
My spouse, who is a native fluent speaker of the languages, follows a large number of Dari and Urdu language Facebook groups and friends. If you think Americans have a problem with reposting stupid things that can be easily debunked by Snopes, you would not believe the amount of naivete on display daily in those language formats. It's hard to explain unless you've seen it firsthand.
People who are one generation removed from subsistence farming are not less intelligent than others. However, they and the family have had a lot less time to view mass media bullshit (tv cable news, internet, etc) and develop the mental callouses necessary to understand the nature of media biases, who owns what media organizations. And in general have had less formal secondary education at the post high school level that teaches critical thinking, skepticism and gain some degree of innoculation against propaganda.
And on a scale of 1 to 10, since I doubt he knows of the existence of advanced video fakery techniques, how likely do you think it is that he could be completely fooled by a faked video of a Myanmar domestic celebrity or politician?
I don't know, for this type of fake video, you don't even need a face mapping or a voice mapping implemented by deep learning.
Just take any random Barack Obama video which we have plenty of, use similar male voice, or copy&paste convenient Obama's voice without context, and some people believe it.
It won't fool everybody but that is not that important. Some people still believe it and act accordingly(Pizzagate conspiracy theory) and that's more than enough to achieve the malicious intent.
Off topic, but what is working for Buzzfeed like? I used to hate everything Buzzfeed-branded, but it seems like it's been improving journalistically recently. How are the working conditions? Do you believe in your purpose?
I think people are seeing the risks of people creating fake videos but there's another risk that I think is being overlooked. As convincing fake videos become a reality and articles like this inform the general public of that reality it becomes incredibly easy to dismiss any inconvenient videos as deep fakes. If Mitt Romney had made his 47% comment in the present climate he likely would have been able to skate by saying that was a deep fake, created by political adversaries.
Many people already fool themselves by simply refusing to fact-check anything that doesn't support their biases. If you thought "alternative facts" were bad now, just wait for this tech to hit 100% believability.
The disturbing fact is that this stuff is progressing in near real-time. A year ago these things were still totally unpassable. Now this Obama video could very easily fool a good percentage of people on mobile. I've no doubt it will be imperceptible within another year or two.
What’ll really be interesting is if someone gets deep voice conversion to produce a plausible output, and combines it with this technique. As is, I think it’s still not enough to really fool people, at least with that particular voice actor... But people really tend to trust voices.
Whether the demo is accurate or massaged, I can't say as I've never used the software myself. But if it is anywhere near as good as their demo suggests, it produces output that is more natural than the current deepfakes video outputs (though it also isn't perfect).
As far as I can tell, it's still early days for the audio faking side of things. I'd give it a year or two before it's indistinguishable from real voice. Then we're in trouble.
It has already been done. 20 years ago a research team created a very convincing fake Whoopie Goldberg. Never heard about it since. Anyone's guess who blacked out that technology.
Right, it's just Jordan Peele doing an impression. I agree that a lot of people would immediately notice the difference and get that it's fake, but I expect there's also a significant number of people who wouldn't.
It's not him for sure but I thought it sounded really good. It's only going to get better, too, when we have a good AI or other type of solution to replicate his voice and mannerisms.
That sounds like a really bad Obama impersonator. The video is OK but still obviously fake.
If you look at the progression of film CGI, with its persistent flaws despite massive budgets, there's no reason to suspect that indistinguishable fakes are coming any time soon.
We've been able to fake voice recordings for decades. Doing it again is not the impressive part of the video. It's a tech demo for the video showing how close it is to working. The point of the video is that this was made with Adobe After Effects and FakeApp and 56 hours of rendering time. COTS software on COTS hardware.
You won't even need a good impersonator, however. The same technology behind FakeApp that blends this video together could be used to blend audio together.
This is supposed to look fake and sound fake because you're supposed to see it as fake. What happens when that isn't true anymore?
Sure, the video is obviously fake. But it's also good enough that it's not "obvious" to anyone who doesn't remember Obama's exact face and voice, and it's good enough that it gives plausibility to anyone calling a real video with slightly distorted audio or video a fake!
I'm a little disappointed that it took this long for people to get over the 'OMG celebrity porno' thing about deepfakes to actually start dealing with the real potential consequences. Deepfakes open up an amazingly large list of questions, from the mundane to the truly challenging. That someone will see imagery that fulfills a sexual fantasy is, by far, one of the most trite and harmless possible uses of this technology.
Realtime video synthesis has been around for awhile. I recall reading in 1999 that if you watched the ball drop on New Years Eve, the advertisements on the buildings you see won't be the advertisements that are actually present in real life. They were being replaced on the fly with different ads. Now, it's basically possible to replace people. Everywhere that people appear. And this isn't something we can deal with simply by going off of gut reaction and intuition, it really raises significant questions.
Who will be the first politician to use a younger version of themselves to promote themselves? Which celebrity will be first to replace not the face, but their body, in film to make themselves more attractive? Which movie studio will be first to film a movie using cheap performers and then re-use the right they have to a major celebrities likeness by just faking them into it? Or maybe they'll only bring in the actual celebrity for close shots?
This can be done on commodity hardware, and will only get easier. It will be used for schoolyard bullying, for amazingly uplifting and important artistic works, for debasement and aggrandizement. It's one of these things that we've got to do some thinking about as a society.
> Which movie studio will be first to film a movie using cheap performers and then re-use the right they have to a major celebrities likeness by just faking them into it?
I believe that studios have done with dead actors. Paul Walker and the guy who played Grand Moff Tarkin in Rogue One being examples that came immediately to mind.
That's the rub, isn't it? Who's a "trusted" new source? All the major players have shown biases, corruption, and manipulation. Lots of people therefore consider some fly-by-night facebook feed that says what the majors don't to be more "trusted" in response.
We shouldn't "rely" on single points of perceived trust, especially in this massive for-profit, for-power media industry. We need to be broadly informed from various angles to have a better chance of piecing together a reasonable sense of what's going on.