What do you think about Veo 3? I mean, it’s really scary how you can generate videos with other people’s faces. This could lead to potential scams, such as fake endorsements and might end up doing more harm than good.
I’m as concerned as you are, Lizzy. I think the capability of Veo 3 itself is impressive. But then, its potential for misuse is massive. I’ve seen people use it to create vlogs of Bible stories and I find it funny. I even made a post about someone using it to create a fictional rich old woman. But then, I’m also scared that people might use it to create stuff related to political disinformation. Like during the war between Iran and Israel, I thought of the fact that someone could create a video of an Iranian commander saying they won’t cease fire after Trump initiated a cease fire. I just think we’re heading toward a future where “seeing is believing” won’t hold up anymore. And without regulation and digital literacy, deep fake could really cause more harm than good.
I have seen many of these kinds of scams as well as content farms. In the UK there is a very popular and well trusted person when it comes to financial expertise, that being Martin Lewis, also known as the money saving expert[1]. He has been commonly used as a way to scam people out of money due to the inherent trust the UK has with him, as seen by this example here as well as in the image below:
These scams are very effective in the UK as in the BBC article here. This is in spite of the fact that Martin Lewis Never does any advertisements in the UK[2].
It is clear that these AI deepfakes are being used as scams and they do work, and that
Is becoming a greater concern, you have to be vigilant about these, as they are used all of the time, AI deepfakes and AI misinformation is already out there and is being spread.
Official Website:https://www.moneysavingexpert.com/ ↩︎
Official disclaimer:https://www.moneysavingexpert.com/shopping/fake-martin-lewis-ads/ ↩︎
Since you opened with that, let’s start with Veo 3, and its introduction was quite the defining, watershed moment in video AI when that threshold was crossed to actual, high-quality video (and the inclusion of sound was just an awesome one-two punch to competitors) from which it’s like the timeline split from BV3 (“Before Veo 3”) to (“After Veo 3”). And to think that in Sam Altman parlance, that was the “worst” that model would ever be, and it only gets better from there.
Sure enough this leaderhip position lasted a few days before the other model providers came out of the gates swinging (and a few coming out of the woodwork as well that you never heard of before) with their own awesome new models (Krea, Hailuo, Higgsfield, Midjourney, and so many others). Each one with their own unique advantage to offer. Midjourney’s was just totally awesome not just in quality but in price.
However, your point is about its potential for misuse, and as you can imagine, these things have already been going on for a while long before Veo 3 (remember those times fake Brad Pitt or fake Elon Musk scammed millions from people thinking they were having an illicit affair with them), except that these make it much easier and of better chances. Which is true, and only gets worse.
The only process to follow is the way these things have always been addressed, which is better education, constant coverage (not just the press but general channels like pods and the like), prompting constant vigilance, and prompting awareness. And to teach people that the best way to combat these things is not to try to know their methods (because there are always new technologies and methods), but to generalize skepticism of wild/weird approaches and news like these, to verify first, especially when weird news of a celebrity is on the news or a celebrity or your relative is keeping you close to eventually ask you for some moneyl
It’s already started happening, unfortunately. With a sharp eye, one can still determine whether or not an AI made the video or not, but eventually it’ll be indistinguishable. On social media apps, like Instagram, there’s usually something like “hey, I’m [billionaire], and you’d better buy my cryptocurrency at [sketchy URL] so that you can make big money”. What’s nice is that there’s a call to action, which generally implies that someone is trying to make money off of the user. Whether it’s ethical or a scam, it’s hard to distinguish at times, and will become tougher to do so.