SoylentNews is people
SoylentNews
Brace Yourself for the 2024 Deepfake Election:
Artificial intelligence was once something the average person described in the abstract. They had no tactile relationship with it that they were aware of, even if their devices were often utilizing it. That's all changed over the past year as people have started to engage with AI programs like OpenAI's DALL-E and ChatGPT, and the technology is rapidly advancing.
As AI is democratized, democracy itself is falling under new pressures. There will likely be many exciting ways it will be deployed, but it may also start to distort reality and could become a major threat to the 2024 presidential election if AI-generated audio, images, and videos of candidates proliferate. The line between what's real and what's fake could start to blur significantly more than it already has in an age of rampant disinformation.
"We've seen pretty dramatic shifts in the landscape when it comes to generative tools—particularly in the last year," says Henry Ajder, an independent AI expert. "I think the scale of content we're now seeing being produced is directly related to that dramatic opening up of accessibility."
It's not a question of whether AI-generated content is going to start playing a role in politics, because it's already happening. AI-generated images and videos featuring president Joe Biden and Donald Trump have started spreading around the internet. Republicans recently used AI to generate an attack ad against Biden. The question is, what will happen when anyone can open their laptop and, with minimal effort, quickly create a convincing deepfake of a politician?
There are plenty of ways to generate AI images from text, such as DALL-E, Midjourney, and Stable Diffusion. It's easy to generate a clone of someone's voice with an AI program like the one offered by ElevenLabs. Convincing deepfake videos are still difficult to produce, but Ajder says that might not be the case within a year or so.
"To create a really high-quality deepfake still requires a fair degree of expertise, as well as post-production expertise to touch up the output the AI generates," Ajder says. "Video is really the next frontier in generative AI."
Some deepfakes of political figures have emerged in recent years, such as one of Ukrainian president Volodymyr Zelenskyy telling his troops to surrender that was released last year. Once the technology has advanced more, which may not take long considering how quickly other forms of generative AI are advancing, more of these types of videos may appear as they become more convincing and easier to produce.
"I don't think there's a website where you can say, 'Create me a video of Joe Biden saying X.' That doesn't exist, but it will," says Hany Farid, a professor at UC Berkeley's School of Information. "It's just a matter of time. People are already working on text-to-video."
That includes companies like Runway, Google, and Meta. Once one company releases a high-quality version of a text-to-video generative AI tool, we may see many others quickly release their own versions, as we did after ChatGPT was released. Farid says that nobody wants to get "left behind," so these companies tend to just release what they have as soon as they can.
"It consistently amazes me that in the physical world, when we release products there are really stringent guidelines," Farid says. "You can't release a product and hope it doesn't kill your customer. But with software, we're like, 'This doesn't really work, but let's see what happens when we release it to billions of people.'"
If we start to see a significant number of deepfakes spreading during the election, it's easy to imagine someone like Donald Trump sharing this kind of content on social media and claiming it's real. A deepfake of President Biden saying something disqualifying could come out shortly before the election, and many people might never find out it was AI-generated. Research has consistently shown, after all, that fake news spreads further than real news.
Even if deepfakes don't become ubiquitous before the 2024 election, which is still 18 months away, the mere fact that this kind of content can be created could affect the election. Knowing that fraudulent images, audio, and video can be created relatively easily could make people distrust the legitimate material they come across.
"In some respects, deepfakes and generative AI don't even need to be involved in the election for them to still cause disruption, because now the well has been poisoned with this idea that anything could be fake," says Ajder. "That provides a really useful excuse if something inconvenient comes out featuring you. You can dismiss it as fake."
(Score: 4, Insightful) by Tork on Friday April 28 2023, @08:41PM (5 children)
I see a couple of things happening in the not-too-distant-future- First is a suspect's video/photographic evidence is thrown out because AI is regenerating an image based on what it's learned landing in a false positive. Second, and this is a good deal more extreme, all smartphone-based photos are no longer considered to be compelling evidence because of lack-of-confidence with the image reconstruction model.
I think you're right that they're going to address it in the way you described but I fear people will find ways around it. If I were the betting sort I'd put $100 on film cameras making a big big comeback. I don't mean as the storage medium, rather a backup for verification. I wouldn't even be surprised if digital artists start using printers for a similar reason. ermm... Im getting fanciful and babbly so I'll just end this post here. :D Man I get bored when my computer takes 10 mins to finish a task.
🏳️🌈 Proud Ally 🏳️🌈
(Score: 2) by JoeMerchant on Friday April 28 2023, @09:15PM (3 children)
I doubt chemical film is going to have any sort of major resurgence. The lowest cost for consumers was around $1 per print back in the day when you factored in cost of film, development and printing - that's like $5 per photo or more today, compared with 0.0005 cents per photo basically in depreciation cost of the phone and maybe 0.0001 cents for electricity.
For people who care (how many people use PGP e-mail or any other kinds of signatures today? Like 0.1%?) it would be simple enough to have an "app for that" which signs you in with fingerprint or swipe pattern or password or whatever and any combination of the above, generating a basis of trust that "it is me taking this video" and then signing that video - both a signature every 30 frames or so, and an overall signature on the entire clip so that you can edit it later and still have the original recording signatures on the smaller chunks, and also put your signature on the editing work.
I have known how to do all this since in practice the 1990s, studied it in school in the 1980s, and, yet, the only application I have ever made of it since then is signing software updates for field distribution for work.
🌻🌻 [google.com]
(Score: 2) by JoeMerchant on Saturday April 29 2023, @01:45AM
Having said all that, our new neighbors with four kids who just paid $1M for a house that was appraised at $385k back 10 years ago when we moved in, their middle daughter has some sort of polaroid contraption that prints photos immediately after taking them. I haven't looked at it closely, but the styling is anything but retro, kinda looks like a purse with holographic prints on it.
🌻🌻 [google.com]
(Score: 2) by lentilla on Saturday April 29 2023, @02:15PM (1 child)
No, it is far from simple. That's the sad truth about security. Common sense tells us it should be simple, unfortunately it really isn't. If it were simple, we would all be using PGP already.
In order to use security (in any meaningful way) the users first have to actually understand the underlying principles - which requires actually thinking about the problem. That's a non-starter for most. Then you have to ignore everybody who tells you they have an "easy" solution (because they don't... otherwise we'd all be using it already).
Once you sufficiently understand the problem, you're ready to go. Now you get to deal with everybody else who doesn't understand what or why you are doing this, and listen to the endless repeats of "you should use this [insecure system], it's much easier!" and "no, that's too hard, it's common sense that it should be easy!".
Then people loose their keyfile and/or forget their passphrase, and get upset because there is no backdoor.
Not that I am suggesting you haven't got the right idea, it's just the implementation is impossible without users actually understanding the problem domain. And users won't spend the time to learn because the siren call of "easy" lulls them into the false expectation that just because they can imagine an easy solution that reality should match accordingly.
(Score: 2) by JoeMerchant on Saturday April 29 2023, @02:44PM
Implementing the cryptography is simple enough, managing the keys is virtually impossible for the bulk of the population, as demonstrated by so many people who lost their bitcoin because they lost the keys.
The cryptocurrency ecosystem sort of demonstrates how digital security does, and doesn't, work for "real" people. Most are using key escrow agents who provide those virtual back doors for when the average customer loses a key - but you've got to trust them like you trust a bank, and being unregulated.... Sam Bankman-Fried (ironic name IMO) and similar cases are to be expected.
There's not really anything the general population takes more seriously than money, and even money isn't serious enough for most people to do security right - so, yeah, e-mails and video "evidence" are taken on a whole lot of faith and not much more.
You can educate people all you want about how to keep a secret, there's still going to be a significant number who will seriously hurt themselves by screwing it up when they protect something important (like their retirement savings) with that secret.
Verbal contracts are still legally binding, even if they're all but impossible to prove in court. The complexities of building trust other ways than a single all powerful secret key (because most people can't handle an all powerful secret key) are what keep these systems from being simple, easily built, maintained and understood. We still use verbal contracts, backed up with recording, written signature, witnesses, various proofs of identity, and threats of legal prosecution for fraud in all those systems. If we're ever going to truly be a global society of billions of people interacting one to one with each other, powerful secrets seem to be the best way to make that practical, instead of the many costly layers of imperfect trust that have evolved over the centuries.
My prediction is that we will slowly evolve toward the one to one use of powerful secrets, those who learn how to use them safely and effectively will benefit from the improved efficiency, and those who don't will be paying layers of trust agents on top of the strong secret users. First in money, but video journalism is another kind of currency that has similar needs...
🌻🌻 [google.com]
(Score: 2) by mhajicek on Saturday April 29 2023, @09:13AM
A good thought, but if I show you a photo print, how do you know the original image source wasn't a high res digital fake?
The spacelike surfaces of time foliations can have a cusp at the surface of discontinuity. - P. Hajicek