Deepfakes have reached a point where, even if you’re looking for tells, it’s possible to miss that a video you’re watching was generated by AI. Early on, the videos lacked coherency, so you’d see things like doors popping open in ways that cannot happen, boxes staying shut even after they’re opened, bugs flying oddly, items not following the laws of physics, et cetera. Now, they’re so good that even if you are actively looking, you may not be able to tell what specifically gives a video away as AI, just that it makes you a little unsettled when you’re trying to watch closely. Even then – not all of them are bad enough to trigger that.
So you’re now trapped in a hall of mirrors, and your loved ones on Facebook have been sharing deepfake videos about things that don’t or won’t happen, and you’re stuck in the middle trying to figure out how you can convince them.
How do you navigate an internet that is actively trying to lie to you, to your face? How do you keep loved ones safe from scams?
The first step to anti-scam measures is being aware that the scam can even happen. For a long time, people were not aware that the bank wouldn’t call them, or that you cannot pay a bill with gift cards, so they fell for those scams. However, once made aware that there’s a scam that runs like that, they stop falling for it. Take some time to explain to them what AI does, and that they can’t just trust a video at face value. For a long time, photos were basically as good as proof that something had happened, until Photoshop made lying with pictures possible; if you explain this is like Photoshop, but for videos, they may grasp the idea that fake videos are basically everywhere now. It doesn’t help that the videos are stupid easy to make!
Alert them that scammers can steal voice libraries and use them to sound like you. For a long time, scammers would call the elderly and try to tell them that they’re a grandchild in danger, who needed money wired somewhere. This was such a huge problem that money-sending apps go out of their way to flash warnings on screen telling senders to double-check the recipient is who they say they are. Crucially, this scam needed someone who sounded like the kid already. If the target could hear well enough to tell it wasn’t their grandkid, the scam might fall apart.
Now, they can steal voices and plug those voices into an AI that speaks for them. This makes detecting a fake harder – but not impossible. Even so, it’s probably better just to establish that you will NEVER call them out of the blue begging for money, or go traveling without letting them know (a common scam is that you travelled to a foreign country and got stopped, and need X-hundred dollars to get out of trouble). If you think you may some day need to call and ask for money because you DID get stuck, set up a code word – if the person on the other side doesn’t know the code word, then they don’t get the money.
On the flip side, even being aware a deepfake video IS fake doesn’t block out the negative feelings it may create. The technology is being used to be annoying and inflammatory, and by golly are the people using it like that succeeding! Facebook reels, if you’re a fan of “justice served” type videos, are full of AI generated videos of porch pirates (people who steal packages off of front porches) picking up boxes with paint bombs in them that go off in their hands. The videos are fake – but for a second, you may have the danger loop in your brain pinged by the thought of someone taking something off of YOUR porch, and what YOU would do about it. Politically motivated videos designed to make you hate your neighbors do a lot of the same. Even if you know it’s fake, you may think to yourself, “why would anyone make this video?” or “why are there so many angry-emoji reacts on this video?”
Or, maybe, you’re trying to watch videos of animals at the zoo – and one of an otter leaping out of its cage to interact with a visitor comes across your feed. You know it’s fake. You know otters are actually pretty vicious predators that are more than capable of seriously harming someone, which is why they don’t have them in petting zoos, and why governments make them difficult to have as pets. But the comments are all cooing about how cute and wonderful it is. You’re left wondering if these are the people trying to walk up to bison to pet them, because they saw an AI video that made a wild animal look like a completely harmless little fluffball and not a 600-lb animal that’s used to fending off wolves. It feels bad. You know it’s fake, but trying to alert the other people in the comments that it’s fake feels like an uphill battle, especially when the groups posting these videos are producing a flood of content. Engaging with it at all boosts their numbers. Meanwhile, that flood is also drowning out responses from real zoos and animal welfare orgs trying to save people from contracting diseases not seen since the 1800s because they thought they’d just casually pet a wild gopher. The flip is also true – you may see some hack for plumbing, or carpentry, or something else you don’t know a ton about, and not realize it’s fake untilan actual plumber tells you that you’ve permanently damaged your sink’s ceramic and you’ll need to drop some real money to get it replaced. Why would the people making the videos care? They got your views. Literally nothing else matters. Harm that befalls you or your loved ones because of their content is not their problem. That’s why they make the content. It is, in actual definition, anti-social. It would hurt less if they actively hated the people they’re misleading, but they literally do not care.
The new internet is leading people down rabbit holes that present real danger to their health and safety if they can’t figure out they’re being lied to, and the lying machines are getting better and better. At some point, it may be better to simply opt out of social medias entirely – it has to be better than getting sucked into videos designed to generate negative emotions or steal your money. Call your family. Check in with them. Make sure they know the videos are easy to fake and that audio clips are getting scary good. And remember that the internet has never accurately reflected reality, so if you need to disconnect, you can.
