With half a moment's reflection you'll realize that photos and video having been showing untrue things since the beginning (before anyone alive was born, in fact) and people have somehow been coping.
In fact, people have always been able to lie, deceive, trick, hide, bullshit, distract and beguile though all of human existence. The technology hasn't changing anything. (Vice seems like the source of the dumbest stuff that gets on to the front page HN.)
I don't agree with this. Sure, things like VFX have existed to augment the experience of movies and other types of content over the years, but these were large-scale commercial projects that only a few could afford.
We're now reaching the point where anyone can create video content that humans aren't able to discern is real or not, and that is unprecedented. Video is an extremely powerful medium that has driven entire movements. For example, BLM wouldn't have happened without the world watching a video of a cop publicly murdering George Floyd.
There's something about seeing video evidence of heinous things happening that drives people to act more than simply reading op-eds or listening to interviews, largely because it's more viscerally stimulating.
We can't know until generative AI has proliferated and fully integrated into society to know whether or not it's changing how we think or behave.
>We're now reaching the point where anyone can create video content that humans aren't able to discern is real or not
>We're now reaching the point where anyone can photoshop an image with intent to deceive
>We're now reaching the point where anyone can film a scene with intent to deceive
>We're now reaching the point where anyone can stage a photograph with intent to deceive
>We're now reaching the point where anyone can mass-produce pamphlets with intent to deceive
>We're now reaching the point where anyone can cheaply replicate a signet ring to pose as a member of our House
>We're now reaching the point where anyone can write a letter impersonating the King
>We're now reaching the point where...
It's funny how little humans change. Think about how, for every single one of the above technologies, many people reacted with the same sense of doom and gloom, induced by a novel technology, yet the general was no different, only the particular. Something else interesting, is many of us relish in this doom-mongering. Again, this is not something new either, as reactionary doomsayers and the cults they've spawned have been coming and going for thousands of years.
I find it useful to remind myself that technologies and socio-philosophical trends come and go, yet man remains fundamentally the same. In times of uncertainty, it's comforting, and informs me how to act.
Yep, and stable diffusion can imitate hand written signatures of everyone, no matter how difficult it's to draw it.
I used ChatGPT to write a story of a person who warns about a new knife technology, which produces sharper knives than any other knife technology in history.
It has always been trivially easy for people to lie to each other.
The fact that it used to be hard to lie via one specific medium, like video, and now is easy (well, soon perhaps -- the cited examples show there's still some ways to go there) is more of a return to the normal course.
No it is not. People used to be able to trust what they saw with their own two eyes. Hollywood could come up with some fake magic, but your average joe couldn’t. Have a look at the new unreal engine demo. Pretty soon, anyone will be able to take a person’s social media photos and videos and place them into any scene they want.
Did you see the news article of the pope in a white coat? Heads up, it was an AI photo and it tricked millions. What do you think might happen if someone decides to flood the internet with a thousand fake George Floyd videos to incite riots because “the ends justify the means”.
You can still trust what you can see with your own two eyes. You can't trust videos or photos of things. Computer image/video editors (e.g. Photoshop, AfterEffects) have made photos & videos untrustworthy for years if a skilled editor made the modifications, and before that some darkroom tricks could do a reasonable job. The new AIs make it cheaper & easier to fake photos and videos.
Yes, optical illusions do exist. They're not particularly useful for the sorts of deceptions that edited media can produce though, so they're not much of a threat.
I wasn’t really talking optical illusions but more something along the lines of the film vantage point or just detective work in general. A lot of people can recall the same event and be telling it faithfully to what they saw and remember but the actual truth of what happened can be something entirely different to their own individual perception. I suppose something like saying you saw someone murder someone with a gun and then run off, when if you’d have seen it from another angle then you would have seen the guy was acting in self defence because he was about to get stabbed or something like that
> What do you think might happen if someone decides to flood the internet with a thousand fake George Floyd videos to incite riots because “the ends justify the means”.
People have been flooding the internet with lies all along -- and, indeed, sometimes incite riots with those lies.
I get it: there are new ways to lie now, and some will be more susceptible because they don't pay attention to the moving target of what's possible with deepfake technology. But adding one more way to lie to the existing pile of a thousand ways doesn't change the balance, especially since the effectiveness of a lie doesn't depend on its believability, but on whether it tells a story the listener wants to believe.
George Orwell in his writings about the Spanish Civil War complained bitterly about the recent loss of any sunset of shared objective truth and how cheap local newspapers would write entirely fictional battles, which, to the readers back home would be taken as true.
If you think augmented imagery started with vfx then you would not believe what people were able to do in a darkroom. Deepfakes can be made with an xacto blade.
This guy gets it. Evil people can frame other people for things they didn’t do, creating completely photo realistic videos as evidence. It doesn’t matter if the truth comes out eventually if the pitchforks have already come out to enact revenge and get their drop of blood. And on the flip side, any evidence of their wrongdoing is just “fake AI news”.
I think the point of the article, and it is a very valid point these days, is that it is the ease in which we turn to the pitchforks that is the problem that truly needs to be addressed.
“We need to put our focus on transforming society to be able to deal with this.”
I mean, I personally don't like this take. People believing lies throughout history has caused a massive amount of damage ("looks like it's time to go to war again").
Lets use an analogy. How much stuff have you lit on fire in your life? You can have something very flammable burning and as long as its seperated some distance from other flammable objects you don't have to worry about it torching everything you know and love.
The problems tend to crop up with fire when you're ignorant or complacent (I'll put stupidity in this classification too). That 'little fire' you thought was under control suddenly isn't, and you can't take it back. You can only hope that either by a massive amount of work you can put everything out, or that the flames will find a boundary they cannot spread past.
When it comes to communications the world doesn't have boundaries any longer. The dumb shit someone posts on twitter can be at the top of Everest and the middle of the Pacific ocean less than a second later. This dumb shit can cause riots. Can cause governments to launch weapons. Can cause neighbors to kill each other. The protections we were afforded of messages took a long time to spread in the past are long gone. We're well beyond tricking one person at a time, or even one nation at a time. We can lie, deceive, trick, bullshit, distract, and beguile an entire planet all at once.
> People believing lies throughout history has caused a massive amount of damage ("looks like it's time to go to war again").
"Looks like it's time to go to war again" isn't a lie, it's an opinion. And the case I think you need to make here is that people believing what at the time were intended to be "lies" have caused more damage than people believing things that were spread in the honest belief they were true i.e. that "lies" are both easily defined and distinctly dangerous.
> "Looks like it's time to go to war again" isn't a lie, it's an opinion.
Eh. The Spanish-American war was started because of a likely false assumption. And posters which literally dehumanize the opponent are a form of lie. There was a time, I've read, when some Europeans literally believed Jews had horns.
https://theisraelbible.com/do-jews-have-horns/ (Due to a mistranslation) "In Christian art of the Middle Ages, Moses is depicted wearing horns, most notably in Michelangelo’s statue created from the 16th century. This led to the persistent anti-Semitic belief that all Jews had horns and were in league with the devil."
> And the case I think you need to make here is that people believing what at the time were intended to be "lies" have caused more damage than people believing things that were spread in the honest belief they were true
I think you're nitpicking. Often enough the second person to spread a purposeful lie does so because they believe it to be true.
A ton of manipulated media that people spread as fact is going to be media made for the fun of it, not as purposeful lies.
I think this is completely wrong. It's now at massive scale, and you are much more likely to be exposed to convincing lies in a world where it's harder to understand what's actually going on. It's possible to live your entire life lying and get away with it because of improvements in living standards and the society we live in.
I know quite a few people who are now living in a parallel reality as they've been exposed to specious nonsense and are indeed immersed in falsehoods. And they point to the volume of it as one of the reasons why these things must be true.
> It's now at massive scale, and you are much more likely to be exposed to convincing lies in a world where it's harder to understand what's actually going on.
This premise isn't valid. The fact that there are a lot more messages going around (and we can safely assume a lot more lies in rough proportion) has no bearing in the likelihood that one is exposed to convincing lies.
1) The more lies, the less convincing each one will be. That's how markets work. Therefore, at some saturation of lies, none will be convincing, and your problem becomes one of discovering truths.
2) Entire societies can believe in a single convincing lie.
3) You don't even have to lie to create untruths. If I were going to have a poll test, as a first approximation I'd ask two questions: 1) Was Saddam Hussein one of the people who plotted 9/11, and 2) Did Russia hack voting machines in 2016, changing to totals to a Trump win? They're good examples of untruths that aren't lies; they were never the stated positions of any authority, instead they were intentionally incepted through FUD and vagaries.
So more lies, fewer suckers, fewer lies, more suckers. And you can create lies without telling them.
Imagine someone making the argument that email and messaging platforms don't change anything from when we had ink and quill scribes. Hey it's just sending a written message to someone, how could it make any mentionable difference?
Technology has absolutely changed many things. The anonymity of the internet means that running scams is much easier. The global reach of the internet means that you can reach across borders and play legal games in a way that was much more difficult before. The automation of phone calls and of sending out emails in bulk means that you can try to deceive tens of thousands of people at a time, instead of just one-by-one. "Spam" barely existed before the internet, and it was definitely much more pervasive after - and the shape and economics of the "scam funnel" is a fundamental part of it.
Sure, "people have always been able to lie, deceive, trick, hide, bullshit, distract and beguile" - but the means that they use to do so has changed, and claiming that "the technology hasn't changed anything" is objectively false. The means matter, and more than just for the fact that changing methods means changing strategies to combat those methods.
Scale matters, and technology has massively increased the scale of many things, including the ability to deceive others.
Vice may be arguing the point extremely poorly, but the point itself is a good one, and worthy of consideration.
This a dishonestly weak take. It's like saying people have always communicated with each other, therefore the introduction of the internet hasn't changed anything. The difference of course with AI is scale and access which is completely revolutionized by the new technology.
This is a complete lack of a take. Instead of saying what it is like, say why it's wrong.
Saying that this view is wrong because of another view that you're also not going to refute is a no-op. The internet didn't change the vast majority of things significantly. It changed the things it changed.
If you're going to do this, you have to explain what characteristics of fentanyl are new, and why that makes the history of opium (and heroin) addiction not the most relevant experiences to apply to the fentanyl scandal. That will be a very difficult thing to do. You can't just handwave at the new thing and say it completely changes the game. That's not argument, that's a car commercial.
The difference is in intensity and quantity. Fentanyl vs opium: breathing in a bit of dust from handling a bag of pills is potentially lethal, and mass manufacturing has put tons of it into the market -- compared to opium which took dedicated farming and harvesting practices and was relatively limited in supply, and overdosing was at least very expensive. ML vs photoshop: results that once took experts many hours of work are now available to anybody with a momentary whim; likewise new voice-cloning scams which would have taken professional impersonators hours of contact with their bait-person has been reduced to a quick phone call.
No. Video and photos (in their original forms) do not show untrue things. They show the "first-hand fact" in a particular time and space.
What can happen is people misinterpret these "first-hand facts" with assumptions and priors that generates "second-hand untrue facts" based on them. If you have all the necessary information to trace back from second-hand untrue facts, you can always discover where the assumptions and fallacies were made.
Now what we have is "first-hand untrue facts" from the source, and there is no way to trace back and find such loopholes. It's similar to not having ground truth / have wrong ground truth when doing supervised learning.
It’s like saying the printing press changed nothing! people could write manuscripts all they want. So what if you drove the cost down significantly? It’s people who have the ideas not books.
> So what if you drove the cost down significantly?
I feel like you're saying this like it's a rhetorical question that you don't have to expand upon. "What were the implications of lowering the cost of print" is a question that takes thought to answer, you can't just breathlessly announce that it's changed everything.
I feel like everyone needs a refresher on distinctions and differences. If lightning struck everyone who spoke English and suddenly they spoke French, everything in the English speaking world would change but very little would be different.
as Exupery pointed out in "Land for people" (?) , technology does not abstract/hide us from nature/essence, quite the opposite, it allows/thrusts us further into that essence extremely (where it might not be pretty to be -
be it middle of mountain / desert, or in the thick of social-(dis)connectivity ..)
Whether we can cope with consequences, is another matter..
The obvious retort here is that network effects modify just what bullshit means. And that the technological difference is qualitatively different. We've all heard that back and forth (it's all been done vs. technological modifications in scope / scale) so many times that I'm not sure it bears repeating. Do you have a take on this general topic?
It changes reach and the ability to flood the space a lie so that the truth is the outlier or flood the space with many lies so nobody knows which to believe.
In fact, people have always been able to lie, deceive, trick, hide, bullshit, distract and beguile though all of human existence. The technology hasn't changing anything. (Vice seems like the source of the dumbest stuff that gets on to the front page HN.)