Why Am I Still Identifying Traffic Lights and Bicycles in the Era of AI?
If I can generate an image of Trump berating a llama in a few seconds, I’m pretty sure our AI overlords can tell an automobile from a mailbox
In the past year, the world has struggled to keep up with the advent of AI. From masterpiece-creating to break-up text-writing, AI has changed the way that many people use the internet entirely. There are few industries it hasn’t yet dipped its robotic arms into.
But in spite of it all, one area of life that reliably makes me feel as though I’m trapped inside a 2007 time capsule is the CAPTCHA (Completely Automated Public Turing test to tell Computers and Humans Apart.)
As more and more of us discuss the realistic possibility of Terminator-styled uprisings, for god’s sake why am I still identifying for my computer what stop signs are? As AI verges on sentience, why am I still distinguishing stop signs from crosswalks? Why am I — a human — trying to prove to a robot that I’m not a robot pretending to be human?
On one hand, the failure of reasoning is a little laughable. Turn on the news and you’ll hear more and more talking heads beginning to panic about the leaps and bounds made within the field of AI.
Scammers have become more clever than ever. AI has enabled them to fully automate their grifting. AI can already recreate voices with a level of accuracy great enough to imitate politicians and destabilize world governments.
It can learn the voices of loved ones and embody them in eerie ways.
… It can allow users to write raps in the style of Spongebob Squarepants characters.
And if AI can endow Nickelodeon characters with laudable rap skills, I’m pretty sure it can tell a school bus from a red light.
While AI is busy wrestling with quantum physics, why am I still here trying to prove I’m not a robot by picking out all the squares that contain sidewalks?
We need CAPTCHAs that keep up with our times. We need a new generation of image identification that’s cybernetic-Schwarzenegger proof — a CAPTCHA so advanced it asks you “which of the following boxes contain some semblance of existential dread?” Or to “Pick out which cat looks most desperately as though it’s contemplating the futility of its ninth life.”
With a new generation of CAPTCHAs, we’ll regain superiority over this digital terrain. We’ll be a step ahead the next time an army of computerized automatons attempts to take over the human race.
AI may be smart, but it’s still years away from developing the proper sense of ontological ennui humans have spent years laboring to cultivate. At this rate, I reason that ChatGPT is still a good 18 months away from its first existential crisis!
And in those 18 months — we have time to prepare for these strange battles ahead. We have time to build a CAPTCHA system that soars well beyond the head-banging lows of the dial-up days.
In a world where I can engage in philosophical discourse with my cell phone, the Herculean task of telling apart a pedestrian crossing from a shadow is a little gratuitous. As more and more of us begin arguing with our smart appliances, why do our most advanced technologies still need us to point out bicycles?
Originally published on Medium
Thank you for reading Ben’s Substack. This post is public so feel free to share it.