I've had to prove I'm human, and I've had to prove I'm me, and the good news is I passed both tests. Because of that prowess I have been permitted to engage with the wider world in terms of buying things from merchants, signing up for offers from various institutions and registering for different group opportunities. And I get it. Each of those operations wants to keep at bay the army of bots being fielded by greedy scammers, diabolical political powers and enemy states trying to gain access. After all, we can't let a bunch of derelicts in WherethehellamIstan take down the economy by placing an order for 10 million lawn sprinklers from Amazon.
But as we are reminded on an almost hourly basis, AI is coming for us all, and it works as well for evil as it does for good. And so it's only a matter of time until ChatGPT and its ilk can fake being a person by selecting all the pictures that have bicycles, or sorting the flowers from the bushes. Of course, it can already do that way faster than I can, but its lack of a physical hand to move the little squares around on my screen is a limiting factor. For now.
As to the mechanics of proving I'm me, I still have the edge, at least for the moment. In one grubby hand I have my keyboard, and in the other, my phone. When I type on one it sends a code to the other. Match the two up, and unless Claude is sitting behind me holding a mouse to my head, I can verify that the person in the middle is indeed moi.
Go a little deeper than a simple verification, however, and we start to hit a problem. That's because the older I get, the bigger the data set is that defines me. More experiences, more connections, more stuff in my life, all of it adding up to the picture of who I am. By its very design, AI excels at crawling and extracting bits and pieces from that information. But I am exactly the opposite: as my dossier gets bigger and bigger, I not only have issues accessing that information, I'm lucky if I remember any specific piece of it at all. Yet that was the challenge that confronted me when I needed to move some accounts around. On a theoretical level I think it's great that the bank didn't want to just take my word for it that I am me; on a practical level that's where the rub occurred.
I was presented not with pictures of buses on the street or a texted code to my phone, but a series of queries based on, well, me. Given the public data points that are floating around out there, it's not hard to scrape the internet and come up with a series of challenges which I should have been able to confirm or deny. Certainly, Grok or Poe or Gemini could parse it seconds, scanning the many years of phone numbers, motor vehicle records, and LinkedIn listings that might contain a passing mention of my name. But me? I can barely recall what shirt I wore on Monday, let alone where I lived in 1969.
But that type of thing was asked of me not once, but half a dozen times. "Was the color of a Toyota you once owned brown, green, or blue?" Uhm, it was never really washed, so hard to recall. "Which of these six phone numbers was once yours?" I can barely remember my number now, let alone one from 1982. "Did you know any of these five different people?" Maybe. "Which of the following corporations have you ever been associated with?" Either my dementia is more advanced than I thought, or "none of the above" is actually the correct answer to that one.
Thankfully, I have retained just enough knowledge about myself (at least at this point) to pass the exam. Going forward, however, I will have to study up before I contact the bank again, and go through old photo albums and yearbooks to jog my memory. That said, if they really wanted to confirm it was me, they should as ask me the stuff that swirls much more readily in my head: song lyrics, joke punchlines, movie one-liners: "Have fun storming the castle!"
-END-
Marc Wollin of Bedford is sure he is who he thinks he is, he just can't prove it. His column appears weekly via email and online on Substack and Blogspot as well as Facebook, LinkedIn and X.
