Cunyan Ma’s Interactive Map Shows Who Builds the Everyday AI Experience

PROVIDENCE, R.I. — Artificial intelligence now shapes searches, recommendations, translations and content feeds. The experience can feel automatic, but the systems are also maintained through human judgment. Much of that work is low-paid, dispersed, and located in the Global South rather than in the offices of the companies whose products users interact with. A public interactive map tries to make that hidden arrangement harder to ignore.

Cunyan Ma, a Brown University undergraduate, has been building a tool that links technology firms, labeling intermediaries, and data-worker locations. Ma works with Brown University’s Center for Technological Responsibility, Reimagination and Redesign. The project is part of a broader effort to examine the physical and economic foundations of digital systems. The map is freely available online. It combines existing databases, Ma’s own collected data, and news reports describing working conditions, so users can trace connections across a supply chain that is often opaque.

Ma said:

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 User Interface to Human Labor

When an image model learns to identify objects, when a language model learns parts of speech in sentences, when a system judges whether chatbot replies are good or bad, and when a medical scan is marked for tumor boundaries, human work is often involved first. Ma notes that the tasks can be tedious. Some also have darker dimensions: harmful material often has to be recognized by reviewers before a system can learn to avoid or filter it.

The demand for these labels can come from social media platforms, technology firms, retailers and defense contractors. The work is often routed through specialist labeling companies, and Ma observes that many of those contractors look for inexpensive labor.

Ma said:

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 arrangement as AI colonialism: difficult, low-paid labor is extracted from the Global South so that profit can accrue in the West.

Inside the Map’s Network

To trace the chain of AI labor, Ma used public databases of labeling firms and their corporate customers. She also built an original database of delivery centers, treating those centers as a proxy for where workers are located. The map connects the dots, and its current form is a partial but suggestive view of the structure.

Network element	Count in the current map	What the element represents
Data labeling companies	13	Specialist firms that receive labeling work
Corporate customers	141	Companies whose products or operations depend on labeled data
Countries containing data worker sites	31	National contexts where worker locations appear in the map

Ma says the map is not a comprehensive view of AI labor chains. The aim is to make patterns more visible, not to present every link.

She said:

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.

More Than an Algorithm Story

Suresh Venkatasubramanian, a Brown computer science professor who leads the Center for Technological Responsibility, Reimagination and Redesign and advises Ma, argues that this kind of visualization can help people see what is hidden by polished interfaces. In that view, AI is not an autonomous presence. It is a product assembled from human knowledge, understanding, and effort.

Venkatasubramanian said:

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.

Where the Map Might Go Next

Ma says she hopes to keep expanding the map and to chart other material foundations beneath algorithms. One planned direction is labor implicated in the semiconductor industry. For Ma, whose concentration is social analysis and research, power mapping is a way to study how global structures shape the distribution of labor.

She also sees another possible function: helping separated workers find one another. The project’s geography can make isolated labor more legible.

Ma said:

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.

Common Misconceptions to Avoid

• AI is purely automated. The map foregrounds the human labeling, reviewing, transcription, and judgment steps that make systems behave in familiar ways.
• Only engineers decide what AI does. Data workers, intermediaries, and corporate customers shape training material and system boundaries.
• The map is a complete inventory. Ma says it is illustrative, not comprehensive.
• Labor conditions are separate from technology design. The reporting and databases suggest that wages, outsourcing routes, and exposure to harmful content are tied to how algorithms improve.

Quick Checklist for Readers

• Identify which AI feature you are thinking about: recommendation, translation, moderation, medical imaging, or another service.
• Ask what kind of human input may have trained it: labels, corrections, ratings, transcription, or judgments about safety.
• Look for the company, contractor, or delivery-center chain if it is shown in public reporting or the map.
• Consider whether the labor is low-wage, outsourced, or concentrated in the Global South.
• Use the freely available online map as an education tool rather than as proof of every supply-chain detail.

FAQ About the Map

Who built the map?

Cunyan Ma, a Brown University undergraduate collaborating with Brown University’s Center for Technological Responsibility, Reimagination and Redesign.

What is the map trying to show?

It connects technology firms and their intermediaries with data workers worldwide, using existing databases, Ma’s own collected data, and news reports.

Which places does it mention?

Ma’s quote identifies Kenya, Indonesia, Colombia and elsewhere as examples of outsourced work in the Global South. The current map records 31 countries containing data worker sites.

How large is the current network?

The map shows links among 13 data labeling companies, 141 corporate customers, and 31 countries containing data worker sites.

Is the map final?

No. Ma says she hopes to keep expanding it and to chart other material foundations beneath algorithms, including labor implicated in the semiconductor industry.

Could it help workers?

Ma suggests a future iteration might help dispersed workers identify one another.

Action Suggestions for General Readers

1. Treat AI as a built system with human components, not a detached intelligent actor.
2. Read the map alongside journalism and worker testimony, because the tool itself addresses the limits of anecdotal reporting while remaining partial.
3. Connect AI features to economic geography: where labor is extracted, where profit accrues, and who bears the exposure to harmful content.
4. Use public education framing when discussing the map with colleagues, students, or readers who need context beyond model performance claims.

For general readers, the practical lesson is simple: when AI feels seamless, it may be because human work has been made invisible. The map gives that work a spatial and corporate context, turning isolated reports into a view of a larger economic structure.