The “ghost work”, a work behind the scenes
As part of a project to develop an application that would help people sort their waste, I had the opportunity to do much needed ghost work.
The “ghost work” what is it ?
Ghost work consists of many humans put at the service of algorithms and artificial intelligence who still need the subtilty of human understanding.
In my case, the project was centered on the labelling and classification of a number of photos of waste products. The end result was to have a database of images from which the AI algorithms can use to decipher one waste product from another. The more the images produced, the higher accuracy the AI algorithm will perform.
My job involved taking as many pictures as possible of many different waste objects. Then via a website called LegIAnnotate, I could sort the images into different categories such as plastic waste, metal packaging, etc.
Thanks to this labelling and classification, the AI algorithm, will be able to classify future images of user’s waste.
Classifying pictures is just an example of one of many tasks occupying the daily lives of these ghost workers. These algorithms need to be fed with data in order to be efficient. We talk a lot about artificial intelligence, but without ghost workers, the AI algorithm would not work. Indeed, for AI to work, it must have been previously trained by humans.
A lot of work for little money
As you will have gathered, ghost work is essential in today’s computerized world. However, executing very simple and quick tasks requiring not much or no skill is extremely poorly paid. Usually ghost work is done by Indians or Americans who would only earn a few cents per job and therefore only a few dollars a day.
Nevertheless, ghost work turns out to be a practical job for some people. Indeed, all tasks can be done from anywhere as long as you have a computer at your disposal.
A great experience
I’m very happy to have contributed to this project. This might seem as repetitive work but I know it’s worth doing. The amount of pictures classified will allow the application to be efficient and help people to better sort their waste. This is another way to make a gesture for the planet.