AI + The New Human Frontier

CHAPTER 16The Rise of Data Colonialism

Much focus right now is on AI's ability to augment human labor and the business efficiencies it will create. But much of this is a “haves” narrative. The effects of AI are not evenly distributed. A darker underbelly supports many of the impacts of AI. AI can enhance and exploit. Augment and abuse. It is creating a new colonial world order, and data workers are bearing the brunt of it.

While tech workers in the developed world may be benefitting from greater freedom and flexibility, these benefits are often built off the backs of tech workers in the developing world. AI is reliant on collecting large amounts of data, but this must be processed and labeled first. This is where the need for a human labor force comes in, which largely consists of underpaid workers in poor countries working grueling hours.

The key to making AI chatbots appear more intelligent and reduce toxicity in their responses lies in a technique known as reinforcement learning from human feedback. This method involves leveraging a dedicated team of human data annotators who meticulously label extensive amounts of raw data to train machine learning models. Data workers in Ethiopia, Eritrea, and Kenya are exploited to scrub online hate and misinformation. Annotators in Kenya were paid less than $2 an hour to filter through piles of disturbing content to make ChatGPT less “toxic.”¹ More than 1,000 images of child sexual abuse have been found in a prominent database used to ...