Naivasha, Kenya – Caroline Njau hails from a family of farmers cultivating maize, wheat, and potatoes in the undulating landscape near Nyahururu, situated 180 kilometers (112 miles) north of Nairobi, the capital city.
However, Njau has opted for a different trajectory in her life.
Presently residing in Naivasha, a picturesque town at the heart of Kenya’s flower industry and positioned midway between Nyahururu and Nairobi, the 30-year-old engages in labeling data for overseas artificial intelligence (AI) firms through a mobile application. Seated in her cozy living room with a cup of milk tea, she meticulously annotates images of roads, intersections, and sidewalks on her smartphone, marking various objects like traffic lights, cars, pedestrians, and signposts. Working on an app designed by an American subcontractor for Silicon Valley enterprises, she earns $3 per hour.
Identified as an annotator, Njau’s data annotation tasks play a pivotal role in training AI systems to identify patterns in real-world scenarios, particularly in the realm of self-driving cars.
“My parents have been hesitant to embrace technology due to the steep learning curve involved. However, my passion for science led me to data annotation, which offers accessible opportunities. You don’t necessarily need a formal degree for this – just a smartphone and an internet connection,” shares Njau, who initially pursued teaching but transitioned to annotation work in 2021.
Kenya is gradually establishing itself as a hub for online endeavors, challenging countries like India and the Philippines. The proliferation of tech startups since the late 2000s, coupled with the influx of tech outsourcing entities, favorable business policies, skilled workforce, and high-speed internet connectivity, has fostered an environment where digital tasks serve as a primary source of income for a significant segment of the youth. A survey conducted by the Kenya Private Sector Alliance (KEPSA) in 2021 revealed that over 1.2 million Kenyans are engaged in online work, predominantly in informal capacities.
Nonetheless, recent accounts from data annotators in Nairobi have shed light on the darker facets of this industry. In a Time article from the previous year, employees at an outsourcing firm in Nairobi recounted the arduous nature of labeling disturbing content sourced from the depths of the internet, all in a bid to enhance OpenAI’s ChatGPT’s ability to identify harmful material. These workers reportedly received less than $2 per hour for their efforts.
In Kenya, a majority of data annotators operate as freelancers, often working from the confines of their homes. For instance, Riziki Ekaka, 45, undertakes data labeling tasks for an American AI corporation from her bedroom, with her young daughter observing while engrossed in play [Anne Kidmose/Al Jazeera].
AI Penetration in Rural Areas
Despite these challenges, the realm of annotation has extended beyond the confines of office spaces in Nairobi.
In a significant development in mid-January, during the inauguration of a government-sponsored tech hub in Kitale by Kenya’s President William Ruto – a town renowned for agriculture near the Ugandan border – a young ICT student shared his experience of earning $284 in three weeks by training AI models for Silicon Valley enterprises. Leveraging platforms like Remotasks, an American website offering remuneration for data labeling tasks, he exemplified the potential for online work.
The video clip from the tech hub, part of a series of initiatives aimed at equipping learners with marketable tech skills, swiftly gained traction on social media, prompting a surge of young Kenyans to register on Remotasks.
“Numerous youth are grappling with unemployment. Even individuals with computer science degrees struggle to secure jobs. The government’s initiative to facilitate access to online work is commendable,” remarks Kennedy Cheruyot, 24, a recent nursing graduate from Eldoret in western Kenya.
Cheruyot, who opened a Remotasks account in 2021, continues to pursue online opportunities while actively seeking nursing positions in healthcare facilities. Several of his peers have transitioned entirely to digital tasks, abandoning their prior vocations.
“In our culture, boys were traditionally tasked with agricultural responsibilities like herding cattle. However, the landscape has evolved, and they are now engaged in online work,” observes Cheruyot during our meeting at a café overlooking Eldoret’s bustling business district, characterized by hardware stores, agricultural suppliers, and vibrant internet cafes, known as “cybers”.
While Cheruyot aspires to own a ranch akin to those depicted in Western movies, his current focus revolves around securing additional online gigs to cover essential expenses such as rent, food, utilities, and transportation.
The surge in commodity prices in Kenya since 2022, attributed to a prolonged drought and the Russia-Ukraine conflict, coupled with the depreciation of the Kenyan shilling due to heightened demand for dollars in the energy and manufacturing sectors, has exacerbated the cost of living for individuals like Cheruyot.
Anticipating potential employment as a nurse, Cheruyot plans to sustain his online engagements, fetching between \(5 and \)20 per hour based on the nature of the task.
“I’m indifferent to the enrichment of Western AI firms through our contributions. As long as we receive compensation, no matter how modest it may seem, it significantly impacts livelihoods in Kenya,” he asserts.
Pioneering a New Wave of Technologists
For Njau, the routine online tasks serve as a stepping stone towards a grander ambition.
“At present, Kenyan annotators tend to someone else’s garden. While the flowers bloom, we remain distant spectators,” she remarks, gesturing towards the lush greenery outside her abode.
“Data annotation is not my final stop; I aspire to progress within the realm of technology. Understanding the trajectory of data and AI programming is my goal. Technology’s pervasive influence necessitates Kenyans to evolve into data scientists,” asserts Njau, who has actively trained individuals with disabilities and young women in data annotation alongside the Nairobi-based non-profit organization, Next Step Foundation. Recently, she secured a scholarship in AI and data science courtesy of the Ministry of Investments, Trade, and Industry.
Initiatives like these strive to position Kenya at the forefront of the technological revolution, as elucidated by Nickson Otieno, the training manager at Next Step Foundation.
“It won’t be surprising if a Kenyan spearheads the next groundbreaking AI innovation. Our youth exhibit remarkable innovation, and a plethora of challenges await solutions. For instance, can AI be leveraged to alert Kenya Power and Lighting Company about power outages by analyzing complaints regarding blackouts posted on social media?” poses Otieno.
Nevertheless, hurdles persist on the path to establishing Kenya – and other African nations – as distinguished AI innovation hubs. According to Professor Tshilidzi Marwala, an esteemed AI scholar from South Africa and the Rector of the United Nations University, a revamp of the education systems is imperative.
“African education often leans towards specialization, a trend observed in countries like Kenya and South Africa with British-oriented educational frameworks. However, in a multidisciplinary world, specialized education falls short,” contends Marwala, underscoring the necessity for a holistic approach. To illustrate, mastery in both medical and computer sciences is essential to develop an AI platform for analyzing x-ray images.
Much of the discourse surrounding AI, exemplified by platforms like OpenAI’s ChatGPT, revolves around the potential displacement of human roles, a concern echoed in African nations. Marwala, however, believes that the apprehensions surrounding AI’s impact have been overstated, often conflating it with conventional automation. Furthermore, AI holds the promise of bolstering small-scale enterprises.
“If a flower grower in South Africa employs AI to assess soil quality via imaging technology instead of relying on a scientist, production costs could plummet, rendering flower cultivation more economical. I envision AI driving enhanced efficiency and cost-effectiveness,” envisions Marwala.
AI applications reliant on data annotated by Kenyan individuals, such as the chatbot ChatGPT, have garnered popularity among tech-savvy youths like Njau and Cheruyot. While Cheruyot acknowledges its utility for accessing recipes or travel plans, he acknowledges that AI cannot supplant the need for human intervention in his professional domain.
