What is implicit crowdsourcing?

The term "crowdsourcing" is a practice that consists of using the creativity, intelligence or know-how of a large number of people (hence the term "crowd") in order to solve a complex task, one that is time-consuming or cannot be automated.

Crowdsourcing should not be confused with user-generated content!

The goal of crowdsourcing is to solve a problem on the scale of an individual, a company or society in general, whereas user-generated content generally only serves to feed the platform in question.

For example, we can talk about crowdsourcing for platforms such as Stackoverflow or Quora, but for social networks such as Facebook or Twitter we would talk about simple content generation.

Implicit crowdsourcing

The term implicit indicates that the user participates in solving a problem without necessarily being aware that their action has just given the system a boost. In contrast, when you report an accident on a GPS navigation app like Waze, you know very well that you are contributing to the life of the app.

1 - ReCaptcha

ReCaptcha is a "classic" captcha service from the user's point of view, who must enter the letters corresponding to the image in order to unlock a feature. But this service, developed at Carnegie Mellon University (Pittsburgh) and later bought by Google, is a true stroke of genius.

Each captcha is made up of 2 words: one that the system already knows and which will allow it to validate that you are human, and another that the system doesn't know and will use your answer to enrich its character recognition.

Thanks to this system, the entire archives of the New York Times were able to be digitized, and until 2014 it also helped enrich Google Books' library.

Since then, Google has repurposed the concept to help it improve its object recognition models in images: when you complete the captcha below, you are therefore participating in improving Google's AI models!

2 - Quick, Draw

Also from Google, the game "Quick, Draw" released in 2016 was presented at the time as a fun demo of their neural network trained to recognize drawings.

Obviously, the objective was above all to push players to feed this network with their own drawings in order to improve sketch recognition with each new game.

Today the intent is announced right from the homepage "Can a neural network learn to recognize doodles? Help teach it by adding your drawings [...]" but when it was released, Quick Draw was the perfect example of implicit crowdsourcing.

3 - Waze

When you're driving and using a navigation app like Waze, you have the option to report hazards on the road, which improves the experience for everyone.

But without you having to do anything, Waze also collects other information, such as your current speed, in order to provide a status report on the level of traffic on a road at a given time (and thus announce/predict traffic jams)!

In 2020, an artist actually simulated a fake traffic jam by pulling a cart containing 99 smartphones down a road (article).

Is it ethical?

From an ethical point of view, the use of implicit crowdsourcing can quickly tip one way or the other.

It's a perfectly respectable trade-off for offering a useful (or fun) service to the user, in exchange for their contribution to a project or application.

The line is surely drawn in whether or not personal data of the user is used, beyond the information they will have provided while using the service.