Finally, a profession that has managed to successfully integrate artificial intelligence into its workflow: Crypto scammers. According to a report from blockchain intelligence firm TRM Labs, crypto scams are up 456% over the last year, due in large part to the ability to produce deepfake audio and video clips with artificial intelligence tools—making good on OpenAI CEO Sam Altman’s recent prediction/warning that a fraud crisis is just around the corner.
There’s no doubt the fraud situation is getting bad. The FBI said it received about 150,000 fraud complaints related to cryptocurrency scams in 2024, with people reporting having lost over $3.9 billion in total. Globally, that figure skyrockets to $10.7 billion according to TRM Labs data. You can go ahead and round those figures way up, too. Speaking to the New York Post, Ari Redbord, the Global Head of Policy at TRM Labs, said that only about 15% of victims actually report these crimes.
These scams are a leveling up of the so-called pig butchering attacks that have become popular in recent years because they don’t just take advantage of people via text. AI now allows scammers to create realistic-looking and sounding audio and video that can trick a person into thinking they are talking to someone real—potentially even a loved one or familiar face. TRM Labs warned that as AI models gain agentic abilities that allow them to interface with things like email and other apps, the process of scamming is going to get automated and a lot more prevalent.
Last week, Sam Altman started ringing the alarm bells on the same problem—though he’s not just worried about scammers taking advantage, but rather the entirety of our existing security apparatus getting defeated. While speaking at a banking regulatory conference, Altman said that AI has already “fully defeated” most authentication services that humans rely on to verify their identity and access their sensitive accounts.
“Society has to deal with this problem more generally,” Altman said, presumably while dressed in a hot dog suit and shouting, “We’re all trying to find the guy who did this.” To that end, Altman’s own company announced earlier this month that it was releasing a ChatGPT Agent that could effectively interact with a computer the same way a human can, switching between apps and completing multi-step tasks that require doing things like logging into different accounts and making decisions.
Altman’s warning of a scam apocalypse seems to have shades of the general warnings that AI execs have been offering about the potential risk of artificial general intelligence, ie, “This could be really bad, but we are absolutely not going to stop.”