Estimated reading time: 3 minutes

‘Deepfake’ Components

Folks in my age group tend to be the targets of text phishing, so when a friend recently shared a screenshot of a peculiar text conversation, it illustrated to me a growing trend in subterfuge. If you’re not familiar, these schemes start similarly with an invitation to some activity, followed by an apology for the “wrong number” and an offer to become friends, because “you seem like a nice person.”

In this example, after starting in English, the phisher inadvertently sends a follow-up text in the Chinese alphabet (translated as “Bella, I'm Evelyn. Are you not in Oregon now?”). Oops. That “technical difficulty” gives them away. Text phishing is a simple, brute-force ploy, and there are other little details which expose the sender as an imposter:

• Crater Lake National Park is truly spectacular, but at over 6,000 feet, the road to the park is closed for winter by November.
• “Bella, I’m Evelyn” is a giveaway. If this truly were a case of mistaken identity, the texter would have been much more informal.
• I can’t remember the last time I reached out to a friend or family member by manually putting their number. “Misdialing” just doesn’t happen like it used to.

It’s hard to talk about fakes and counterfeits nowadays without mentioning artificial intelligence. Mainstream media is awash in stories about what generative AI can do, and not always for the greater good of mankind. I recently read about a speech synthesis company claiming it could create a complete model of someone’s voice from as few as 45 key phrases. It wouldn’t take much audio to “steal” our voices, and it’s no wonder that we see a wave of deepfakes of government leaders and political figures across social media containing spoofed voices.

Of course, doctoring information to deceive is nothing new, but the level of sophistication now possible makes it increasingly difficult to distinguish real from counterfeit. Attention to subtle details is the key to detecting and avoiding a counterfeiter, a deepfake, or a phishing text.

So, what does all this mean to the EMS industry? Counterfeiting has been an ongoing threat for the last 30 years or so, and one can argue that it’s on the increase. Our reputation as assemblers is at stake if we lose control of our parts inventories and find ourselves building product with counterfeit parts. While it might sometimes be as easy as paying attention to the component labeling, counterfeiting techniques now include functional spoofing of the genuine part—the EMS equivalent of a deepfake video, perhaps? Our vigilance and sophistication must rise to match the inventiveness of the bad actors. That is what inspired this edition of SMT007 Magazine.

In this month's issue of SMT007 Magazine,  Paul Jarski, whose real-world experiences set the stage for our other conversations on counterfeiting. We then follow up with watchdog organizations like CALCE and ERAI, who share how they help prevent and/or mitigate counterfeits in your supply chain. We also contacted Tony Bryant, who developed (and conducts) a course on counterfeiting for IPC. In his article, Tony takes a deep dive into his counterfeit mitigation course syllabus to share how to respond when counterfeits are found or suspected. We also go back to a column by Bill Cardoso, who explains why you should “X-ray everything.” 

ERAI indicates that counterfeits are shifting to meet market shortages, maybe, but not letting up. In this world of artificial intelligence, we all must be on the lookout for deepfakes meant to persuade and, unfortunately, to deceive. So, whether it’s a phishing scheme in your text messages, sociopolitical propaganda in social media, or counterfeiters in your supply chain, taking the necessary precautions will help keep you above the fray.

This column originally appeared in the October 2024 issue of SMT007 Magazine.