TikTok ran a deepfake ad of an AI MrBeast hawking iPhones for $2 — and it’s the ‘tip of the iceberg’::As AI spreads, it brings new challenges for influencers like MrBeast and platforms like TikTok aiming to police unauthorized advertising.
Simpsons called it:
Oh boy. This is all moving very quickly. People already fall for simple SMS scams, I can only imagine just how many more will be falling victim to this trash in months/years to come.
People have already been falling for scams that “Elon Musk” was promoting. Naturally I’m talking about these crypto schemes run by scammers on YouTube using a deepfake of Musk. It’s been happening for about two years now.
Bill Gates has been giving away his fortune to some lucky email recipients every year now since the days when you had to pay for the internet by the hour.
Don’t even need AI video to do that. People just give away tens of thousands of dollars based on a shitty email, written in bad English, with Bill Gates’ name on it.
Just imagine fans getting a facetime call from a Taylor Swift, explaining they won half-price tickets to an exlusive fan event. Then “Taylor” has to drop out to make the other calls, but will leave them a link for the purchase - only valid for 15 minutes, as of course many others are waiting for this opportunity.
This is the entire basis of using an adblocker like ublock origin. It is purely defensive. You don’t know what an advertising (malvertising) network will deliver, and neither does the website you’re on (Tiktok, Google, Yahoo, eBay, etc etc etc). With generative AI and video ads and the lack of content checking on the advertising network this will just get worse and worse. I mean, why spend money on preventing this? The targeted ads and user data collection is where the money’s at, baby!
Related note, installing uBO on my dad’s PC some 8 years ago was far more effective than any kind of virus scanner or whatever. Allowing commerce on the Internet was a mistake. That’s the root of all this bullshit, anyway.
Fuck ads in general. I don’t care if they are legitimate or not I don’t want to be mentally assaulted every time I try to browse a website.
That too.
Allowing commerce
on the Internetwas a mistake. That’s the root of all this bullshit, anyway.That’s more accurate.
You can’t put everything back in Pandora’s Box. Right or wrong this is the world we live in. What is lacking is regulation. If left to their own devices we (royal) are shitheels. Unfortunately we (royal) are persistent shitheels which is why when we put in regulations we then strive to rip them out.
Do you mean “we (humans)”? Because “the royal we” just means “I”. Like how the queen says “we are not amused” when they mean “I don’t like it”. Related to how in many European languages (including early-modern and older English) the plural is a polite form of address (like tu and vous in French, du and sie in German, thou and you in English)
Honestly ads aren’t that bad when done right. For instance, the yellow pages had ads but they didn’t follow you wherever you went. We need ads that people want to look at. If I’m trying to find something g to buy I don’t mind looking at an ad. I just don’t want it to be everywhere and end up being malious
Although I’d say a good 99% of all ads are terrible, I have yet to find any that are absolutely egregious when visiting sites like FurAffinity. Totally depends on the site and what ad services they chose or are forced into if they have ads. Also where and they are placed.
I never thought of it in security terms but that makes sense. If nothing else, having it installed makes for a better web experience.
The more I hear about AI-generated content and other crap that is posted online these days, I wonder if I should just start reading books instead, maybe even learn to play on a musical instrument and leave virtual world altogether.
Plagiarized books entirely written by an AI are nor far-fetched. Get ready for a shitty reality.
Why would AI need to write a plagiarized book? Isn’t that just a knock off?
Because writing a knockoff book takes time and effort, and getting an AI to do it does not.
The same question applies to many fields. Why get AI to write a plagiarized tweet or Reddit comment? It’s a lot less effort than a book, yet both platforms are plagued with bot comments.
There is already books written entirely by AI too, AI is litterly getting everywhere
I’m sure it’s also writing the music too.
And some forum comments.
Beep boop
And some forum comments.
Good thing concert tickets aren’t overpriced and only available via resale, so I can see humans perform human music.
Keep informed so you don’t buy AI-written books. Better know what’s going on instead of letting the world happen to you imo.
Butlerian jihad sounds like a good idea rn
Yeah but then I need to hire a mentat to keep my shit straight
No.
We need the AIs to make the Men of Gold so we can compete with the murder orgy space elves.
Came to watch a fake mrbeast left dissatisfied. Came back to post a link: https://twitter.com/MrBeast/status/1709046466629554577
Click blewlowaugh now
And that is why we need a pixel poisoner but for videos.
I’m not familiar with the term, and Google shows nothing that makes sense in context. Can you explain the concept?
Here specifically it’s a technique to alter images that makes them distorted for the “perception” by generative neural networks and unusable as training data but still recognizable to a human.
The general term is https://en.wikipedia.org/wiki/Adversarial_machine_learning#Data_poisoning
One example of a tool that does this is https://glaze.cs.uchicago.edu/ but I have doubts about its imperceptibility
Yeah I’m at a loss aswell. Is it a way to prove the source of a video?
Its AI poison. You alter the data in such a way that the image is unchanged to a humans visual eye, but when imaging AI software uses the image within its sample imaging, the alterations ruin its ability to make correlations and recognize patterns.
Its toxic for the entire data set too, so it can damage the AI output of most things as long as its within the list of images used to train the AI.
That seems about as effective as those No-AI pictures artists like to pretend will poison AI data sets. A few pixels isn’t going to fool AI, and anything more than that is going to look like a real image was AI-generated, ironically.
It can seem like whatever you want it to, its already been used and has poisoned data sets.
Wake me up when orgs like Stability AI or Open AI removed about this technology. As it stands now, it’s not even worth mentioning, and people are freely generating whatever pictures, models, deepfakes, etc. that they want.
It’s a bit unclear what you’re after here. Don’t do it unless it’s already perfect?
Why would they openly removed about it? Thats free advertising that it works. Not to mention, you cant poison food someone already ate. They already have full sets of scrubbed data they can revert to if they add a batch thats been poisoned. They just need to be cautious about newly added data.
Its not worth mentioning if you dont understand the tech, sure. But for people who make content that is publicly viewable, this is pretty important.
It’s sort of like the captcha things. A human brain can recognize photos of crosswalks or bikes or whatever but it’s really hard to train a bot to do that. This is similar but in video format.
Can Mr Beast sue til tok over this?
No, can only DMCA takedowns.
If he can find the creator or publisher, he can sue for misappropriation of name and likeness. It’s a privacy tort.
However I think he already said he doesn’t care
Section 230 will absolutely reign here.
No, Section 230 protects TikTok as a platform. He would have to sue the ad creator.
This is the best summary I could come up with:
TikTok ran an advertisement featuring an AI-generated deepfake version of MrBeast claiming to give out iPhone 15s for $2 as part of a 10,000 phone giveaway.
The sponsored video, which Insider viewed on the app on Monday, looked official as it included MrBeast’s logo and a blue check mark next to his name.
Two days ago, Tom Hanks posted a warning to fans about a promotional video hawking a dental plan that featured an unapproved AI version of himself.
“Realism, efficiency, and accessibility or democratization means that now this is essentially in the hands of everyday people,” Henry Ajder, an academic researcher and expert in generative AI and deepfakes, told Insider.
Not all AI-generated ad content featuring celebrities is inherently bad, as a recent campaign coordinated between Lionel Messi and Lay’s demonstrates.
“If someone releases an AI-generated advert without disclosure, even if it’s perfectly benign, I still think that should be labeled and should be positioned to an audience in the way that they can understand,” Ajder said.
The original article contains 518 words, the summary contains 168 words. Saved 68%. I’m a bot and I’m open source!
I read this title last night and thought it was a story about AI making an amalgamation of MrBeast and Stephen Hawking to shill iPhones
I mean, that’s why I clicked
It would be a good time to start a paid online community that has a lengthy vetting process for accuracy and authenticity of all content.
Haha, paid, good one!
It’s either paid by the users or by the advertisers, or do you have a better plan on how to fund running such platform?
Nope, I don’t, otherwise I’d do it because it would be a massive change to whats out there.
I was a bit too short on my comment I admit. I am laughing about the suggestion because making a completely new community with paid access from the start will doom it immediately. It needs to have a certain pull to get users to actually pay something and so far I’d say thats either failed or platforms have not quite taken off.
Same as converting a free community to a paid access model, for that you just need to look at what happens to Twitter. Its a slow bleed but it certainly damaged the platform. Instagram is next it seems with the ad-free tiering that will no doubt lead to obnoxious ads. We will see if they achieved the necessary pull to keep people there or not.
Yes, because SomethingAwful has struggled over the decades so much.
Ok, they have like 200.000 users registered. I know its historically an important site but in terms of userbase that is nothing.
And do they have a lengthy vetting process for accuracy and authenticity in place as you are asking for?
Who said this platform would be the biggest?
You’re making a lot of assumptions here, and it’s coming off as seal-lioning so I’m going to disengage from you here. Cheers.
Well, the thing is that only big platforms struggle with the problems of the topic we discuss here. Smaller platforms just do not have the userbase size where it would be viable to try and scam them for money. Which is why this happens on tiktok in the example up here.
So I just don’t agree with your viewpoint and you don’t with mine. I take no offense to that and I hope you did not either, that was definitely not my intention.
Yes, paid
Guess who wouldn’t be in the target demographic?
Moderation is costly if your dedicated to only putting out accurate and truthful content. And people will pay for that accuracy when the media landscape becomes saturated with AI deep fakes.
Link to video?
Paywalled. :(
Ban TikTok already. Can someone tell me why this isn’t a good idea?
So you see this as specifically a tiktok problem and not a tech problem? Do you think it won’t/hasn’t happened elsewhere, and will be only a tiktok problem? I don’t use tiktok, or care about it but I feel like every problem with it is something endemic to social media platforms run by businesses atm.
Sorry, just blurting an admittedly off-topic idea. No, I don’t think TikTok is the problem here…
But why isn’t TikTok banned? It is cancer.
So is all social media. People were saying the same about Facebook. People are saying the same about Twitter. People are also saying the same about reddit. TikTok is just the main popular right now.
People were right about facebook and twitter, and arent even that wrong about reddit.
I dunno man, maybe social media aint a great idea
You’re posting on a social media platform…
A non-centralized platform though
Maybe it’s a bad idea for the government to completely shutdown a private business?
Maybe it’s because of the first amendment violations the government silencing a social network would entail.
Maybe it’s because we are attempting to be a democracy and have due process available to all entities that operate within the United States?
Just a thought.
Because its a overstep of government power. How would that be enforced? The US couldn’t even manage to get rid of alcohol let alone software.
Government? A ban would happen via the app stores, like all the other mass banned apps.
What about side loading and third party app stores? They would have to patrol peoples devices.
Not to mention its a little concerning that apps can be pulled from the app store for any reason. They bill would allow them to pull any app including ones that protect peoples privacy and freedom.
… what bill? I literally just said there was no government involvement.
You can side load a social media app all you like, if its pulled from the major stores it has lost its lifeblood. And, uh. Most apps can be pulled for nearly any reason, cause there is no government owned public software store. The company store can kinda do as it likes.
This does not end until anonymity on the internet does. As long as nobody signs their content, nobody can be blamed for pooping in the punch bowl.
And that’s why I have conspiracy theories about the conspiracy theories about digital ID.
Removed by mod
A signature that isn’t tied back to a single human being with a name and a face and an address that you can send a court summons to isnt worth the flash memory it’s printed on.
The only difference is having a database of name to public key. The rest would end up even easier to track than crypto trades.
They do all this stuff for money. If it’s $2 for a phone, follow where the $2 goes.
If it’s just about who posts it, they’ll just hack an account and post the shit to some random account without a lot of followers. That’s how the Elon Musk crypto scams on YouTube do it.
