Amazon's detailed customer dossiers are just the tip of the iceberg in the AI era

• This post originally appeared in the Business Insider Today newsletter.
• You can sign up for Business Insider's daily newsletter here.

Boxers or briefs? Don't worry. AI already knows.

A new trend is making people realize the amount of data that tech companies have on them and how they're using it.

Amazon's "About You" feature — an AI-created rundown of your preferences as a shopper — has been making the rounds on the internet. Amazon's specificity has been eye-opening, from "has flat buttocks" to "uses phone in the shower to read Kindle books."

(Here's a step-by-step guide for finding your "About You" feature.)

Ben Shimkus canvassed the BI newsroom to see what Amazon thinks of his colleagues.

Amazon's read on me wasn't all that surprising (or funny), with one exception: "performs manual work or construction." Yup, that's me, punching the clock at the old newsletter factory.

Companies building profiles on us shouldn't be surprising. But the depths they're willing to go might be.

When a Wired reporter recently requested a copy of his customer data from McDonald's loyalty program, he got back a 515-page dossier. And this wasn't necessarily someone you'd consider a frequent customer. McDonald's predicted he'd visit 2.16 times over a six-week period.

McDonald's data collection is small fries when you consider other companies you're interacting with.

AI giants are using their agents to get embedded deeper into users' personal lives. Instead of companies having to do some legwork to collect info on customers, people are actively giving them the keys to the data castle.

From email, calendars, and texts to bank accounts, it's all fair game. And for those who don't want to have to think about reordering paper towels, it's worth it.

But even if you want to overlook the security or privacy concerns, there's still reason to give at least a slight pause.

AI companies are under pressure to generate revenue from the tools they're spending billions of dollars building. A subscription model is the most straightforward play, but that risks excluding a good chunk of users who won't be willing to cough up.

Tapping into the troves of data they've collected on customers is another route. Suddenly it's not just ordering ahead for you, but ushering you toward certain brands.