There was a natural awe that came with the rise of ChatGPT.
AI's infrastructure foundations date back to the mid-20th century. However, the adoption we’ve seen over the last few years feels like the greatest leap since the dot-com boom. In fact, corporations are actively weaving AI usage into employee performance reviews! But beneath the AI productivity promises are structures that can look less like innovation and more like digital colonialism. In Empire of AI, Karen Hao provides insights into these structures within countries like Kenya and Venezuela. Her book reveals how technology ecosystems may be sacrificing human dignity in pursuit of rapid innovation.

The $2 AI infrastructure double standard

There’s a profound economic hypocrisy within tech “disaster capitalism” that Hao exposes in ChatGPT's development.
OpenAI outsourced grueling content-labeling (and graphic) work to cheap labor markets in the Global South, where economic conditions made this work difficult to refuse. However, under the pressure of the workload and psychologically damaging tasks, these workers did exactly what Silicon Valley tells the public to do. They used AI to automate their workload.
On the surface, OpenAI’s decision to stop using Kenyan workers who were automating aspects of the process made sense. If ChatGPT trains ChatGPT, it defeats the purpose of human oversight. But the real lesson here isn’t about the quality of the algorithm. It's about the wellbeing of the people behind it.
Innovations that prioritize the bottom line over the survival and wellbeing of workers only prove that this AI Empire didn’t want intelligent tech operators. What's more, these multi-billion-dollar valuations are perhaps structurally impossible without the massive pool of workers earning under $2 an hour. The 1% of the 1% aren't just building brilliant code; they are scaling on the backs of human exploitation.

AI infrastructure: Innovation at the cost of compassion

Ultimately, true innovation should prioritize individual human dignity, transparency, trust, while valuing people’s time (my 4 T’s). But instead, the current AI infrastructure treats human beings as things to squeeze, discard or blacklist, for the sake of margins and convenience.
This isn't just an overseas supply-chain issue. It's a canary in the coal mine for how the tech industry views human labor as a whole. So next time we type a prompt into AI like ChatGPT and marvel at its efficiency, it's worth asking what and who made that experience possible. We must be willing to look under the Silicon Valley red carpet that tech giants have laid.
Until we value the humanity of the workers at the baseline as much as the code at the top, we aren't building the future, but just scaling an exploitative empire.

Get insights like this delivered monthly.

Subscribe to the CancerGeek newsletter for exclusive perspectives on healthcare innovation, digital health trends, and more delivered straight to your inbox.