You've been free training artificial intelligence for Google for 15 years, just kept in the dark the whole time.

You’ve been training Google’s AI for 15 years. You had no idea.

By Sharbel

Reprinted from Mars Finance

Every day, about 500,000 hours of human labor are used for free by Google. And those contributing are just trying to log into their online banking.

reCAPTCHA is the most successful invisible data operation in internet history. At its peak, 200 million people completed verification daily. But almost no one realizes what each click really means.

Google’s self-driving car company Waymo is now valued at $45 billion. Most of its core training data is provided for free by you when you visit various websites.

Here’s the full story:

Origin: A Clever Idea

In 2000, spam bots were destroying the internet. Forums were flooded, inboxes were overwhelmed, and websites desperately needed a way to distinguish humans from machines.

Professor Luis von Ahn from Carnegie Mellon University solved this problem. He invented CAPTCHA: distorted text that only humans can read, but bots cannot.

But von Ahn saw more than that. Millions of people spent effort on these challenges. What if that effort could do two things at once?

In 2007, he launched reCAPTCHA. Its cleverness was that it no longer displayed random gibberish, but two words: one known to the system, and another from real scanned books that computers couldn’t recognize yet. Your responses helped digitize these books.

These books come from the archives of The New York Times and Google Books, totaling up to 130 million volumes.

You thought you were just logging into a normal website, but in fact, you were helping OCR (Optical Character Recognition) for the world’s largest digital library.

In 2009, Google officially acquired reCAPTCHA.

Later, Google changed the game

The era of “distorted text” ended around 2012.

Google faced a new challenge: Street View cars captured every road worldwide, but the photos were raw data. To make AI effective, it needed to understand what it saw: street signs, crosswalks, traffic lights, storefronts.

So Google redesigned reCAPTCHA v2. Instead of distorted text, it used image grids. “Click all the squares with traffic lights.” “Select every crosswalk.” “Identify storefronts.”

These images came directly from Google Street View. Your clicks became labels.

Every choice told Google’s computer vision models: this cluster of pixels is a traffic light, that shape is a crosswalk. You weren’t just passing a test—you were building a dataset.

Unimaginably Large Scale

At its peak, 200 million reCAPTCHAs were solved daily. Each challenge took about 10 seconds, meaning 2 billion seconds of human effort per day—that’s 500,000 hours.

Paid data annotation costs roughly $10 to $50 per hour. Using the lowest estimate, the value of free labor extracted daily is up to $5 million.

And reCAPTCHA isn’t just in one app. It’s everywhere—banks, government portals, e-commerce sites. You have no choice: want to log in? First, label some data. Google never asked your opinion, paid you nothing, and never even told you about it.

What has all this created?

These data directly feed into two products:

• Google Maps: the world’s most used navigation tool. Its ability to recognize street signs, shops, and city geography owes part of its success to billions of human labels during login.

• Waymo: Google’s autonomous vehicle project. For safe navigation, self-driving cars need to recognize thousands of visual patterns nearly perfectly.

The true training data for these recognition tasks was labeled unknowingly by millions through reCAPTCHA. Waymo completed over 4 million paid trips in 2024, valued at $45 billion. Its foundation was laid by those “unpaid internet citizens” just trying to check their emails.

Why can’t anyone replicate this model?

Data annotation is extremely expensive. Companies like Scale AI, Appen, and Labelbox exist to solve this problem—they hire hundreds of thousands of workers, sometimes earning less than $1 an hour.

Google’s solution is different: they make annotation mandatory. No payment, no consent—just a “ticket” to access every corner of the internet. The result: billions of labeled images, global coverage, all weather conditions, every city in the world. No annotation company can do this. The internet itself is a factory, and every netizen is an uncontracted worker.

You are still participating

In 2018, reCAPTCHA v3 was introduced, which no longer shows challenges. It observes your mouse movements, scrolling speed, dwell time. Your behavioral fingerprint tells it whether you’re human. These behavioral data are also fed back into Google’s AI systems.

You never actively opted in, there’s no checkbox to agree. But on most websites you visit, you’re still doing this.

A Disturbing Irony

Luis von Ahn’s original idea was brilliant: turn the effort humans already waste into useful output. But what Google has done with this vision is another story. They exploited a security mechanism users had to use, deploying it across the web, harvesting data to build billion-dollar products. Users gain nothing, and are often unaware.

The deepest irony is: you spend years proving you’re human by completing visual recognition tasks that AI couldn’t do at the time. But once AI masters these tasks, human visual labeling is no longer needed.

You proved you’re human, only to make yourself replaceable.