A dad, husband, author, and journalist living in New York decided to find out. For 48 hours, A.J. Jacobs would avoid all interactions with A.I. and machine learning.
He woke, picked up his phone, and entered his iPhone passcode like it was 2017. He quickly learned his iPhone would be useless. No AI curated news, social feeds, or attention maximization targeted ads. No email passed through the spam filter or podcasts cleaned up with AI. He put his iPhone in the drawer.
His wife Julie turned on the lights and A.J. quickly flicked them off. “Are you kidding me?” she asked.
Con Edison uses AI to monitor four million meters to manage the grid. He thought about using rainwater to brush his teeth after realizing New York water uses machine learning to monitor 1,600 sensors. He had to walk or bike to avoid traffic flow monitoring, Ubers, and subway. He couldn’t use weather apps, Zoom that leverages AI for noise suppression, credit card transactions, food services, retail, or television streaming. He ended up watching Brewster McCloud on a twenty-year-old DVD player.
AI and machine learning (ML) models are used to predict, detect, and recognize patterns. The outputs of AI and ML models feed explicitly programmed software scripts that essentially run our world. While people are most familiar with AI through Chatbots, like ChatGPT introduced by OpenAI in November 2022, the hidden AI in the background runs our world. It has been since the 1990s when it began detecting credit card fraud, managing retail inventory, and sorting zip codes for the U.S. Mail.
The fraud detection ML model has milliseconds after a credit card is swiped to either flag for review (and anger the customer) or clear (and risk bank losses). These ML models consistently improved over the decades to find the perfect balance of customer friction and costly losses. Kroger would make five million ML model sales predictions per day, one for each item in each store, to power their supply chain. There’s not enough room in the back for extra Cheerios boxes or tolerance for unhappy Cheerios lovers.
Jacobs wrote in his New York Times story, “What I didn’t expect was that my attempt to avoid all interactions with A.I. and machine learning would affect nearly every part of my life — what I ate, what I wore, how I got around.”
Most can go forty-eight hours without chatbots, few can go without hidden AI and ML models that power human society. Jacobs mentioned that even a goat herder in the mountains check weather apps.
Photo Source: The New York Times