ItInsightsWorld Feature: Inside Cunyan Ma’s Brown Map of Global South Labor Powering AI

Modern AI systems are often described as autonomous software, yet the services that recommend posts, route traffic, and surface media are also dependent on paid human review and labeling. PROVIDENCE, R.I. [Brown University] is the setting for a project that turns this dispersed labor into a visible economic structure: Cunyan Ma, a Brown University undergraduate working with the Center for Technological Responsibility, Reimagination and Redesign, has built an interactive map of the firms, contractors, and work sites that keep many algorithms operational in the Global South and beyond.

The tool is freely available and gathers existing data-worker databases, information Ma assembled, and news reports describing conditions many workers face. Its purpose is to make the human supply chain behind machine output easier to see.

Ma frames the project’s intent directly:

My primary motivation is to visualize the idea of AI colonialism in an impactful and immediately accessible manner

The gap I’m trying to fill is that most of what we know about data workers comes from anecdotal reports from authors and journalists. Those are great because they give us a very detailed look at working conditions and pay, etc. But I’m hoping this might start to fill in a larger picture of what AI colonialism actually looks like.

From chatbot training to hazardous image review

Training a model can begin with simple tasks: tagging objects in photographs, marking parts of speech, transcribing lengthy speech, rating whether chatbot responses are good or bad, or outlining tumor boundaries in medical scans. Ma notes that some work is even darker, because systems must first recognize disturbing content in order to filter or prevent it. Data workers may therefore have to view horrifying images or text passages so labeling can teach models what should be restricted.

Many corporations — including social media companies, technology firms, retailers, and defense companies — send data labeling to specialist companies, which then seek inexpensive labor.

Ma summarizes the geographic consequence:

How that manifests in reality is that companies outsource that work to the Global South — Kenya, Indonesia, Colombia and elsewhere

So you have folks around the world who have to read paragraph after paragraph of AI-produced child pornography, bestiality, violent content or whatever else, just to try to train AI to be safe.

Ma describes this as AI colonialism: difficult, low-paid labor is taken from the Global South so that profit can accrue in the West.

Mapping firms, contractors, and worker sites

Ma traces the chain deliberately kept opaque, using existing databases of labeling firms and corporate customers plus her own database of delivery centers as proxies for worker locations. The current map links 13 data labeling companies, 141 corporate customers, and 31 countries containing data worker sites. Ma says it is not a comprehensive view, but she hopes it illustrates the structure.

She describes a public education purpose:

One purpose for this is as a public education tool

When people want to learn about AI and its social impact, the map puts it in the context of this larger economic and political structure.

The humans beneath the algorithms

Suresh Venkatasubramanian, a Brown computer science professor who leads the Center for Technological Responsibility, Reimagination and Redesign and advises Ma, argues that such visualization helps people see AI as an aggregation of human knowledge, understanding, and effort.

It’s all humans underneath

That’s the real story behind the generation of AI. The problem with a lot of current discourse around AI is that it seeks to ignore that. The result is that we start constructing stories around AI as its own sentient being, as something that’s alive. But it’s not. It’s something that’s built — something that’s constructed by us, by people, all across the world.

What the map may reveal next

Ma hopes to expand the map to other material foundations of algorithms, including semiconductor labor, and to keep using power mapping to understand how global structures shape labor distribution. She also sees a possible practical role for worker connection.

One of the things about this is that it’s distributed

People don't even know about their fellow co-workers. That makes unionizing or anything like that just kind of impossible. Maybe they can use a future iteration of the map to identify each other.

For ItInsightsWorld readers, the article’s core point is that the project shifts the conversation from abstract AI capability to a documented chain of companies, contractors, and countries, making visible who is asked to handle the least pleasant work and where the resulting value is collected.