In our last post we explored the process the New York Times went through to determine whether the photos Donald Trump posted of Nicolás Maduro were real. We dug into how normal people can follow a similar process to check incoming media. In this post we’re going to take a look at how various actors took advantage of making synthetic media for “fun and profit”. We’ll view this through the lense of the information vacuum around Maduros capture.
The “Fun” Part of Synthetic Media for “Fun and Profit”
According to the Straits Times, the creator of one of the most widely reposted FAKE images of Maduro in American Custody was created by an “AI video art enthusiast” who goes by the name “Ian Weber”. Here’s how he did it.
- First he went to Google’s Gemini image generator.
- Second he gave it a simple prompt.
- He cropped out the Gemini watermark and posted the result on X.
- He got lucky with the timing as his post was within 20 minutes of Trump’s post.
- Despite him only have 78 followers the image went viral.
“Fun” right? Please note thick sarcasm if you haven’t already.
The “Profit” Part of Synthetic Media for “Fun and Profit”
Deloitte’s Center for Financial Services predicts that gen AI could enable fraud losses to reach US$40 billion in the United States by 2027. This is up from US$12.3 billion in 2023. That’s a compound annual growth rate of 32%. There’s plenty of money to be made in deception.
Probably we’re only seeing the tip of the iceberg in terms of what’s going to happen in the future. In the case of the breaking news with Maduro’s capture, within minutes of the news breaking the web was flooded with fake Maduro pictures. Those pictures captured millions of views on social media. In an “attention economy” those views got paid out as cold hard cash by platforms like X. Our CEO recently penned an article about how influence has become currency this has resulted in the truth becoming a casualty. It’s well worth a read.
Moreover, this was seized upon either knowingly or unknowingly, by public figures to serve their own narratives. For example Portugal’s far-right Chega party significantly amplified the reach of the fake image created by “Ian Weber” via their official account. A polish MP followed suit on his twitter account.
Euronews did a deep dive after the fact about all the various synthetic media flooding our streams. It wasn’t just fake images of Maduro. It was also fake images of crowds cheering the event, fake images venezuelans mourning and even a fake image of President Hugo Chavez’s mausoleum as a bombed out wreck. We’ll delve into the consequences of this of politically motivated synthetic media in a future post. Needless to say the profusion of hyperrealistic videos is no bueno for the body politic.
Existing Safeguards Aren’t Enough
After the storm of images subsided the new york times did an investigation. Gemini and models from OpenAI and X all readily and easily produced these kinds of deepfakes with very little friction. Watermarking like Google’s SynthID are a good start but far from sufficient. All in all, the genie is out of the bottle and there’s likely no putting it back inside.
Confused? Turn to vali.now
Next time you’re overwhelmed by uncertainty about what to believe you can reach out to the cybersecurity professionals at vali.now. We don’t pretend to know everything but we deal with this stuff full time. Regardless of the context we can lend our expertise and experience to help you parse the fakes and identify reality with our Scam Shield Services.