5 ms·
These deep fake articles are becoming a meme. They mostly seem alarmist, and yet they're not authored by people actually in the industry. Deep fakes automate w
by echelon 7y ago
These deep fake articles are becoming a meme. They mostly seem alarmist, and yet they're not authored by people actually in the industry.
Deep fakes automate what deep pockets and state actors could already do with Photoshop and other professional tools. The world isn't going to become a scary place because the barrier to entry got lower and the technology has been democratized. People are smart. Fakes will be detectable through entropy measures, corroboration, common sense, etc.
FWIW, I've been working on real time voice to voice style transfer.
https://drive.google.com/open?id=1zRvJEGJjTpKvvzel-J0agh3fKBn9aqGy https://drive.google.com/open?id=1zRvJEGJjTpKvvzel-J0agh3fKB...
There are already a few other (non-real time) players in this field.
I'm hoping to spin this up as a small social app or filter and sell it so I can fund my capital-intensive film making startup.
I think this tech should be widely available. Not only will it make people think and question more, but it'll be fun too.
It's also amusing (and terrifying) to see all the anti-1st Amendment legislation aimed at combating deep fakes. The truth is that there is nothing to fear except our freedoms being taken away.
- ipython 7y agoI don’t think you should dismiss these concerns so quickly. It sounds like you have experience in this field. Perhaps that would make it easier for you to spot potential fakes? What about your grandma? How would she fare? And besides, the end game isn’t to fool everyone into believing a fake. No, the more insidious goal is to flood the zone with enough dis- and misinformation to overload our ability to filter it. It’s like gaslighting at scale- at some point you just stop being able to process information because it’s so voluminous and of dubious quality that you stop believing any of it.
- qball 7y ago>No, the more insidious goal is to flood the zone with enough dis- and misinformation to overload our ability to filter it. It’s like gaslighting at scale- at some point you just stop being able to process information because it’s so voluminous and of dubious quality that you stop believing any of it. This isn't new by any stretch of the imagination. It might as well outright be the job of the Fourth Estate ever since the Fourth Estate has existed (funded, as always, by ads). And when you look at the people complaining about this, which is for the most part people part of said political group (Facebook/Twitter/non-tech journalists), it makes sense that anything that could diminish their ability to control the narrative should be presented to the public as a new and novel threat. So this isn't going to result in widespread social chaos, because it's already happening. If anything, this technology will democratize what the old media is currently doing; which will likely result in driving political power towards crowdsourced content aggregation instead of the individual content creators themselves.
- EGreg 7y agoI would go further. Think of it like an ELO rating where most grandmas are 1200, and you pride yourself on being 2000. But eventually, the fakes become so good that the best humans can’t distinguish them, and half the time neither can ANY system, except with some probability. What then? An unlimited amount of glut that is indistinguishable from human writing, but can be produced with any kind of bias, by any number of fake identities and there is no reasoning with hem.
- nitwit005 7y agoThere have been photographic fakes for a long time now, and they were quite good even in the pre-digital era. Have you seen it creating major social issues? I haven't observed much of anything. For the lies most people want to tell, there isn't any need to make a fake. Want to paint a protest in a bad light? Just use a real photo of a riot and falsely label it as the current protest.
- kevingadd 7y agoWith how readily fake/modified content proliferates on the internet already and how much money tabloids make off made-up stories and faked photos, can you really argue that there is zero negative social impact from fakes? Even for someone with decent media literacy, proving 'this is faked' is a lot harder than 'this is some old content falsely labeled as new content', because you can't just look for the original source. You have to try to identify what the original source was pre-fake, and that's a LOT harder with deepfakes.
- nitwit005 7y agoI didn't argue there was zero negative impact. I just said it wasn't major. If the tabloids were unable to make fake photos, would it affect anything at all? They can just write up a fake article without a photo. The actual act of making fake media just isn't necessary for any of this behavior.
- andreilys 7y agoPizza Gate led to an armed individual shooting up a pizza store because he thought it was an undercover pedo ring. So yeah, I imagine the fabrication, proliferation and automated spreading of conspiracy theories to have potential major societal impact
- nitwit005 7y agoYou're arguing against some sort of straw man argument I didn't make, but I'm not entirely sure what it is.
- bransonf 7y ago> What about your grandma? How would she fare? Grandma’s still falling for the phone and mail scams. No amount of legislation is going to fix the reality of technological illiteracy among the oldest adults. Deepfakes might fool some of today’s adults who don’t quite understand, but we are raising a generation that has turned into a meme: “Everything you read on the internet is true” -Abraham Lincoln I think the real silver lining here is that the internet is an alternate reality. Many of us refuse to believe that, but social media has created manufactured people. The only solution is to bring people back to the real world. The people are real here. Their opinion, no matter how controversial, comes from a real mouth, and the face you see is the one they were born with. If anyone forms their worldview based entirely on things they read on the internet, they probably would be just as susceptible to our real world forms of propaganda/gaslighting/ whatever you want to call it.
- skybrian 7y agoThis essentially means the web is too difficult for some users and they need something else, like maybe an app store. Maybe some company will win big by providing a safer (or apparently safer) alternative? Previous examples: Gmail had a better spam filter. Apple and Google did a better (though not perfect) job of protecting users from arbitrary code execution, as did the web itself, way back when. This doesn't happen all that often, but if it succeeds, power users will scoff at how nerfed the new thing is. I'm reminded of an old story [1] about an early game for children: > I found myself unable to reconcile the idea of a virtual world, where kids would run around, play with objects, and chat with each other without someone saying or doing something that might upset another. Even in 1996, we knew that text-filters are no good at solving this kind of problem, so I asked for a clarification: "I’m confused. What standard should we use to decide if a message would be a problem for Disney?" > The response was one I will never forget: "Disney’s standard is quite clear: No kid will be harassed, even if they don’t know they are being harassed." But maybe text filters will be better if you throw enough machine learning at the problem? [1] http://habitatchronicles.com/2007/03/the-untold-history-of-toontowns-speedchat-or-blockchattm-from-disney-finally-arrives/ http://habitatchronicles.com/2007/03/the-untold-history-of-t...
- 7y ago
- blunte 7y agoI strongly disagree that this lower barrier of entry won't make things significantly worse. I also disagree that "people are smart". Of course that depends on your definition of smart, but evidence of general intelligence is evaporating rapidly. If through software a non-technical person can generate hundreds of websites of "content" that either pollutes a topic with noise or misinformation, you can bet that will be used at great scale for everything. Pick an element of human life, and I can describe a way someone will have financial or political (which often boils down to financial) motivation to do it.
- cm11 7y agoThis reminds me of Ze Frank's Ugly MySpace pages. He used to tease creators of "ugly" MySpace pages during his show. When someone called him out on it, he had this to say about it: "For a very long time, taste and artistic training have been things that only a small number of people have been able to develop. Only a few people could afford to participate in the production of many types of media. Raw materials like pigments were expensive; same with tools like printing presses; even as late as 1963 it cost Charles Peignot over $600,000 to create and cut a single font family. The small number of people who had access to these tools and resources created rules about what was good taste or bad taste. These designers started giving each other awards and the rules they followed became even more specific. All sorts of stuff about grids and sizes and color combinations — lots of stuff that the consumers of this media never consciously noticed. Over the last 20 years, however, the cost of tools related to the authorship of media has plummeted. For very little money, anyone can create and distribute things like newsletters, or videos, or bad-ass tunes about "ugly." Suddenly consumers are learning the language of these authorship tools. The fact that tons of people know names of fonts like Helvetica is weird! And when people start learning something new, they perceive the world around them differently. If you start learning how to play the guitar, suddenly the guitar stands out in all the music you listen to. For example, throughout most of the history of movies, the audience didn't really understand what a craft editing was. Now, as more and more people have access to things like iMovie, they begin to understand the manipulative power of editing. Watching reality TV almost becomes like a game as you try to second-guess how the editor is trying to manipulate you." http://www.zefrank.com/thewiki/the_show:_07-14-06 http://www.zefrank.com/thewiki/the_show:_07-14-06
- tootahe45 7y ago'Deep fakes' are right up there with 'AI is going to turn against us', or 'software engineers are going to be replaced by AI' on my bs list. The real problem we face with elections is much less exciting to talk about. It's a few people going around making reasonable and factually correct comments, and manipulating upvote/downvote counts with bots to drive-home the belief that one view is popular, and another is extremely unpopular.
- luc4sdreyer 7y ago>AI is going to turn against us Would you mind sharing your argument about why this notion is a waste of time? Machine learning experts are concerned about it [1], but maybe you know something they don't? [1] https://www.technologyreview.com/s/602776/yes-we-are-worried-about-the-existential-risk-of-artificial-intelligence/ https://www.technologyreview.com/s/602776/yes-we-are-worried...
- tsbinz 7y agoYou are misreading the article you are citing. Quoting your article, "Contrary to misrepresentations in the media, this risk need not arise from spontaneous malevolent consciousness. Rather, the risk arises from the unpredictability and potential irreversibility of deploying an optimization process more intelligent than the humans who specified its objectives." See paperclip maximizer. Even that is too narrow for the questions they asked IMO - someone who says that the development of an artificial intelligence will be "on balance bad" could be thinking that this AI would be in the hands of the 1% and so further contribute to inequality, even if it doesn't act unpredictably.
- luc4sdreyer 7y agoI completely agree with what you said, except for the part that I misreading the article. GP's comment is vague, it could be anything between "artificial intelligence will be on balance bad", and "the Terminator movie will happen IRL". My question was an attempt to clarify the statement and understand why it's reasonable to dismiss the idea as BS.
- TaylorAlexander 7y agoIt’s interesting that you say this. I suppose I have a different experience. I’ve been following the news around the supposed military coup in Bolivia. And it turns out that thousands of pro-coup bots have appeared on twitter basically overnight: https://mobile.twitter.com/BenjaminNorton/status/1193715544949694464 https://mobile.twitter.com/BenjaminNorton/status/11937155449... It’s making it hard for me to engage with others online. I want to talk to others about my political beliefs but I’ve seen infiltrators like the ones on twitter destroy conversations and spread FUD. I experienced this with Venezuela before. In my opinion it is a real problem. Yes we will eventually adapt but I think actors taking advantage of this tech early will succeed in influencing the public.
- unityByFreedom 7y agobots and shills are different. The idea that someone can make a bot to effectively spread propaganda without being detected is far-fetched. The idea that you can convince someone to write propaganda is real.
- PeterStuer 7y agoThe deeper problem isn't the perfection of fakes. It is the further onslaught of social destruction of perceived/accepted 'truth' as everything always could be fake by the mere acceptance of the potential existence of the technology. It is the bad actors getting of the hook more than the bad actors framing the innocent.
- b0rsuk 7y ago> People are smart. Fakes will be detectable through entropy measures, corroboration, common sense, etc. Yes, in the same way that people can detect and filter email spam. Either by a time-consuming manual effort, or complicated systems very few people understand. And we have corporations to swoop in and save us by offering a centralized service! Search wasn't hard enough to enter. It will be fun in the same way that searching for something useful in a landfill is fun. Even if 5% people click on mail spam, it's still profitable and there will be more of it. If 5% is deceived by a fake video, it can have a big impact on election results. Fake news and outright lies are already damaging enough if you do get a court order to print an apology. By that time it's already damage control, you can't revert it because people move on and believe what they want to believe. On social media people upvote what gut feeling tells them is right.
- WA 7y agoI just read an article from the linked blog (http://thismarketingblogdoesnotexist.com/ http://thismarketingblogdoesnotexist.com/). It takes me way too long to decipher this as gibberish. Far longer than most spam emails.
- LoSboccacc 7y agoI see this as a net positive. as you say people with resources and an agenda have been manipulating the media since, idk, Caesar's deeply exaggerated Commentaries or even before and has been going trough the ages, including Stalin's photoshops and Kerry Fonda revisionism or at a simpler level the billion 'anonymous sources say' articles that litter modern journalism. but now that the tools are getting into the hand of everyone, everyone is starting to questioning everything they see, and it was about time that people start to realize the chain of trust required in reporting and accepting news. here's hoping that in half generation we'll be back at thinking critically about every piece of information we see.
- Braggadocious 7y agoOne key thing here is that the sheer quantity of deepfake content will proliferate exponentially quicker than anything user-generated, and continue to be generated faster, and faster, faster, and faster... on all levels, articles, media, AI will be taught to program and create its own websites and social media and apps etc. So while users may be able to tell what fake is from real (although I doubt it since a decent chunk of the population can't even tell what's fake and real now) the amount of BS everyone will have to sift through online will be insurmountable. You'll have AI sifting through the trash for you, which means your feeds and search rankings will be more, and more contrived and curated with less and less feedback from you or other users. THere was a time when exploration of the internet and discovery of websites was a natural experience. There were no ads giving everyone the same five websites for everything, the same solution for every problem. If people had a question, or needed a solution for something, they had to find it themselves and so the function of social media, message boards, google, etc was that everyone on the internet was in the process of finding their own way, their own solutions. So you'd find crazy inventive solutions and websites for things. That's been disappearing slowly. Once 99.9999999% of the content on the internet is AI created, that experience will never come back. It begs the question: what is the innate value of the internet and what would have to be lost for everyone to stop using it? Maybe someone will create an amish version of the internet whereby every user must be verified as human by another human who was verified by another human etc. before they can log on, and everything on it was user generated, and user programmed, etc.
- buboard 7y agoMany things can change. On the one hand, deep fake text/video competes with (legacy) spammers, which means the profit from spam will soon drop to zero. OTOH, our trust in search algorithms will drop dramatically. People will need to start signing their public media and announcements, probably using one of the existing and robust publick blockchains. It's quite possible that google will become a lot more useless, and some curation will be needed for search results. Even without the advent of deepfakes it was bound to happen as the internet keeps growing in content. Fake news is just the tip of a big iceberg