Technology entered a new age last year, and ever since the “AI revolution,” all I could hear were people screaming from the rooftops about how “This is the future” and “It will change the industry as we know it.” I’m not one to buy into this sort of hearsay without giving it time, as these claims can quickly collapse in on themselves when no one finds a real use for said “amazing” new technology.
It doesn’t help that a number of the AI demonstrations that I’ve seen are unconvincing. “Look! We can make robotic-sounding NPCs interact with you in a video game!” “We can create uncanny videos and images of ‘photorealistic’ people!” “We can steal people’s art, feed it into our algorithm, and mimic their style!” All of which I translate to mean: “We can avoid paying developers, coders, and actors by having an AI do the work for them.” Even then, the ethically sound demonstrations usually involve automatic framing or background blur in cameras, which isn’t very exciting.
With an onslaught of soulless AI art, The Finals using AI voices instead of just hiring real people, and the film Late Night with the Devil using AI-generated title cards, this “revolutionary” technology is in a pretty dire state. Do I believe AI can help? Absolutely, but it has yet to prove itself. Until it does, I’ve seen no reason to sing its praises.
While I’ve waited patiently, that patience snapped this past weekend.
Authentication processes are getting nightmarish
When AI was first introduced, I noticed an immediate change in the steps websites and applications took to verify you aren’t a robot. These systems had gotten a massive security upgrade. This makes sense on paper, as the growth in capabilities for AI obviously means it can be used for more nefarious deeds. The original Captcha systems are likely too primitive to protect against modern AI.
But the key to that upgrade, weirdly, comes in the form of “fighting fire with fire.” Yes, the irony here is that they’re using AI-generated images to combat AI. Real people are now thrust into a battle between artificial intelligence, forced to prove that they’re not AI to an AI. I noticed this several months ago, but I broke this past weekend when trying to make a new X account.
X showed a selection of six images, each of which had a collection of four dice on them, some with the typical engraved dots to showcase their numbers and others having actual Arabic numerals. My goal was to look at these six images and determine which collection of dice added up to the number 14. While that sounds simple enough, X expects the user to do this ten times. Ten rounds of looking at six images is already a pain in the ass, but it’s especially frustrating when the AI makes a mistake.
During my first time authenticating, I noticed that the dice were sometimes in the middle of tilting over. This is normally fine, as it’s simply just a slight tilt and the number is still visible at the top, but in one of the rounds, the dice was tilting too far away from the screen and it wasn’t visible. I checked all of the images to see if any of them were the correct answer, but they didn’t seem to be, so I clicked the messed up image in hopes that the AI meant to make that one the correct answer. By the end of the ten-round authentication process, I was denied.
I’m not sure if it was that round that denied me, but I then had to redo the whole thing and felt like I was trying to prove myself as human to an AI that couldn’t even handle simple dice generation properly. I will admit, replicating this bug so I could take an image took a dozen or more tries, so I clearly got pretty unlucky, but it happened regardless. Making things worse, sometimes, halfway through the test, it has an error that suggests “Something went wrong” and forces you to start over from the beginning.
Bottom line
I’m tired of dealing with authentication. Captcha was already a pain in the butt, but now? These processes are far more frustrating to deal with, solely because AI is getting smarter. In my day-to-day life, this negative experience is one of the only ways this technology is actually impacting me.
I have no doubts about the potential of AI and how it can improve our livelihood, but I’m starting to get tired of waiting. When will it actually help people?
