People in Venezuela are annotating the endless training data for vehicles' AI.

Data Farm

In order to teach a car to drive itself, its artificial intelligence needs to be presented with thousands of examples of things it might encounter on the road: pedestrians about to cross a street, for instance, or specific street signs.

Now, with economic turmoil in Venezuela that left many impoverished, companies that gather and label this training data for self-driving car developers are relying on cheap labor from the South American country, according to MIT Technology Review.


In 2018, hundreds of thousands of Venezuelans signed on to work for AI data aggregating companies including Mighty AI, Playment, Hive, and Scale, MIT Tech Review reports. The vast scale of the operations illustrates once more how cutting-edge AI tech often depends on the grunt work of impoverished workers.

The pay isn't great, MIT Tech Review reports, but workers likely flocked to these companies in particular because they provide steadier work than other data labeling jobs that tend to ebb and flow over time.

"[The Venezuelans] were aware that, on one level, it's exploitation and they have to do it because everything else failed them," crowdwork expert Florian A. Schmidt told MIT Tech Review.

READ MORE: Unable to buy cars, Venezuelans labor to make sure robotic ones don’t crash [MIT Technology Review]

More on self-driving car technology: Expert Slams Waymo Safety Data: “They Obviously Do Not Trust These Numbers”

Share This Article