UPDATE: the robocallers have definitely gotten more human sounding over the last twelve (!!!) years, but they’re still giving themselves away.
The most sophisticated call I’ve gotten so far started with the caller “talking to a co-worker” with a tone mimicking a person making an aside as heard through a slightly covered (physically muted) microphone. “I’ll take two sugars in my coffee. You’re a blessing!” Followed by a full volume, “Excuse me- hi there! I’m __________ from ___________.”
Etc.
But I did as I usually do when I suspect I’m not talking to a real person: I accused it of being a robot. This always results in a denial, commonly along the lines of “It must be my microphone” or blaming the speakerphone.
But if/when I ask a personal question- like asking for their shoe size- they always reveal themselves. To date, all responses belong to one of two categories:
1) the robot completely glosses over the question and proceeds with its spiel, or
2) it goes silent and/or disconnects, sometimes abruptly, sometimes with “let’s talk about this at a later time.”
The scam scripts are also veering into straight up gaslighting, which pisses me off. Nobody in my household is suffering notable cognitive impairments, but I’ve gotten some calls that would probably snag someone who did.
The most common gaslighting opening has been an Aussie woman’s voice who says I asked her to call me back. That’s something I NEVER do, so the game is over immediately. There’s a few others who make similar mistakes, claiming I asked them to do XYZ, again, things I idiosyncratically do not do EVER.
Relatedly, I’ve also seen people talking about detecting video call scammer weaknesses. One that still happens but is at the tail end of its efficacy is asking the suspected fake to hold up 3+ fingers in front of its face. At one point, that would glitch the image out, revealing the falsehood.
The newest software has gotten good enough that it usually won’t glitch with that test. But because the refinements of the programs occur by accretion, many AIs will still avoid holding up their fingers in front of their faces.
The other noteable flaw is that scammers using facial filters- at least older ones- will do almost anything they can to avoid turning their faces away from the camera. If they do, the filter will cease working, revealing who is being concealed. I watched one recently where the caller, asked to show a closeup on something from the shelf behind them, awkwardly groped to grab the item without once taking their eyes off the camera. The recipient terminated the call.
The most sophisticated call I’ve gotten so far started with the caller “talking to a co-worker” with a tone mimicking a person making an aside as heard through a slightly covered (physically muted) microphone. “I’ll take two sugars in my coffee. You’re a blessing!” Followed by a full volume, “Excuse me- hi there! I’m __________ from ___________.”
Etc.
But I did as I usually do when I suspect I’m not talking to a real person: I accused it of being a robot. This always results in a denial, commonly along the lines of “It must be my microphone” or blaming the speakerphone.
But if/when I ask a personal question- like asking for their shoe size- they always reveal themselves. To date, all responses belong to one of two categories:
1) the robot completely glosses over the question and proceeds with its spiel, or
2) it goes silent and/or disconnects, sometimes abruptly, sometimes with “let’s talk about this at a later time.”
The scam scripts are also veering into straight up gaslighting, which pisses me off. Nobody in my household is suffering notable cognitive impairments, but I’ve gotten some calls that would probably snag someone who did.
The most common gaslighting opening has been an Aussie woman’s voice who says I asked her to call me back. That’s something I NEVER do, so the game is over immediately. There’s a few others who make similar mistakes, claiming I asked them to do XYZ, again, things I idiosyncratically do not do EVER.
Relatedly, I’ve also seen people talking about detecting video call scammer weaknesses. One that still happens but is at the tail end of its efficacy is asking the suspected fake to hold up 3+ fingers in front of its face. At one point, that would glitch the image out, revealing the falsehood.
The newest software has gotten good enough that it usually won’t glitch with that test. But because the refinements of the programs occur by accretion, many AIs will still avoid holding up their fingers in front of their faces.
The other noteable flaw is that scammers using facial filters- at least older ones- will do almost anything they can to avoid turning their faces away from the camera. If they do, the filter will cease working, revealing who is being concealed. I watched one recently where the caller, asked to show a closeup on something from the shelf behind them, awkwardly groped to grab the item without once taking their eyes off the camera. The recipient terminated the call.
