CBx Arrow.png

Hi Fam

Welcome to CultureBanx, where we bring you fresh business news curated for hip hop culture!

Google & Microsoft Banking On Africa’s AI Labeling Workforce

Google & Microsoft Banking On Africa’s AI Labeling Workforce

By CultureBanx Team

  • Kenyan AI data labelers make around $9 a day

  • AI data preparation expected to hit $1.2B by the end of 2023

Information prepared by data labelers in Africa, the construction workers of the digital world, create an important part of Silicon Valley’s efforts in AI. Companies like Google (GOOG -0.64%), Microsoft (MSFT +0.05%), Salesforce (CRM +0.52%) and Yahoo use Samasource, a U.S. firm that creates AI training data and information around images, by using some of the poorest tech laborers in Africa. Since AI is nothing without human labeling, will tech companies continue to walk the fine line of off-shoring what is slated to be the most dynamic part of the global economy?

Samasource Computer Workers Kenya.jpg

Why This Matters: Data image labeling is a necessary part of machine learning and there's an obvious motivation for these companies to use workers in parts of the world where wages are rock bottom. Not to mention the market for AI data preparation was $500 million in 2018 and is expected to hit $1.2 billion by the end of 2023, according to Cognilytica.

Samasource provides what it calls a living wage to Kenyan AI data labelers of around $13-$16 a day, compared to the average worker in the country who makes about $3 a day, but it’s still chump change for Silicon Valley companies. They would be hard pressed to find U.S. data labeling experts to do boring, repetitive, never ending work at that price point. Low wage jobs like these don’t just exist in Africa they are also in Southeast Asia.

Many of these workers often live below the poverty line, and continue to fuel a new type of blue-collar industry around curating the data that powers AI.

The main reason big tech companies farm out these jobs is because data preparation and engineering tasks represent over 80% of the time consumed in most AI and machine learning projects. Since time is money, they don’t want to pay people a lot of it for repetitive labeling.

Situational Awareness: The human in the AI loop isn’t going away anytime soon for data labeling and AI quality control. It’s also going to be a while before the technology these African workers are helping to create will be used in communities where they live. Many of these workers often live below the poverty line, and continue to fuel a new type of blue-collar industry around curating the data that powers AI.

CBx Vibe:Price On My Head” NAV Feat. The Weeknd

Trump vs China: Diving Into the Black Consumer Impact

Trump vs China: Diving Into the Black Consumer Impact

LeBron’s Education Philanthropy Fuels School Choice

LeBron’s Education Philanthropy Fuels School Choice