They're TRAINING YOU

Overview

The video argues that “Are you a human?” CAPTCHA / reCAPTCHA systems are not only designed to block bots, but also (often without meaningful consent, pay, or credit) collect human labor and behavioral data to train AI systems.

Key Points and Claims

• CAPTCHAs are framed as security checks

  • Users are interrupted by tests (e.g., clicking images, reading blurry text) meant to prove they’re human.

• CAPTCHAs also train AI

  • The video claims reCAPTCHA—linked to Google—uses users’ actions to label real-world images (such as traffic lights, buses, and crosswalks), creating data to train machine learning models used for tasks like search and mapping.
  • It notes that reCAPTCHA is associated with Google, referencing Google’s 2009 acquisition.

• Broader AI training beyond reCAPTCHA

  • The video further claims that leaked documentation indicates similar data types helped train Waymo / Google self-driving systems, suggesting users indirectly contribute to autonomous vehicle “brains.”

• Behavioral surveillance beyond the checkbox

  • The video argues that the “I’m not a robot” checkbox and related flows assess how users interact with the site, using behavioral signals such as:
    • mouse movement
    • page context
    • scrolling speed
    • cursor hesitation
    • whether users are logged in

• Ethical and labor concerns

  • The central critique is that these systems use users’ time, attention, and actions as low-cost (or free) input for powerful technologies—possibly including applications the video associates with surveillance, law enforcement, and military—without informed consent.

• Not limited to one company

  • The video claims similar systems exist across the ecosystem, connected to government agencies, facial recognition startups, and data brokers.
  • It describes this as creating an “ethical blind spot” (as characterized by researchers).

Presenters / Contributors

• No specific presenter or contributor names are provided in the subtitles.