By Colbert Gwain in Bamenda, Cameroon
Prior to the popularization of Artificial Intelligence (AI) in Africa, particularly in Cameroon, a significant amount of information about my village, Muteff, in the Fundong Subdivision of the Boyo Division, North West Region, existed online, especially on social media platforms.
When I began exploring the AI world and its functionality, I was confident that searching for information about Muteff would be effortless. However, I was surprised when I first used Meta AI to find information about Muteff, and it responded that it couldn’t find any information on the topic, suggesting that it might be a name, word, or term unfamiliar to it.
The AI asked me to provide more context or clarify what I was looking for, even inquiring if I was searching for a person named “Muteff” and requesting details about their profession or achievements.
When I specified that Muteff is a village in the Fundong Subdivision of the Boyo Division, North West Region, Cameroon, the machine acknowledged the village’s existence but lacked specific information about it. I was advised to search online maps, Cameroon government websites, local organizations, and research papers for more information.
I also inquired about the mystical and revered Nantang Yoh juju in Muteff, but Meta AI stated it was unaware of such a mystical order, although it recognized the term “juju” as referring to traditional medicine, spiritual practices, or supernatural powers in Cameroon. After several hours of back-and-forth interaction, Meta AI eventually acknowledged that Nantang Yoh might be a mystical order in Muteff village.
In response, I explained the significance and importance of Muteff village and the Nantang Yoh mystical order, which I believe was a precursor to conventional and codified artificial intelligence due to its predictive intelligence in community affairs. Meta AI responded that it was simply a machine working with available information.
Furthermore, I asked Meta AI if it was aware that the Ijim mountain forest near Muteff village was home to the Venus Flytrap (Dionaea muscipula), a carnivorous plant known for capturing and digesting insects.
Meta AI categorically stated that Muteff could not be a natural habitat for the Venus Flytrap, suggesting that if found there, it would likely be cultivated, as the plant is indigenous to North and South Carolina in the United States, not Africa.
Such realizations, particularly about the paucity of information on Africa in AI systems and, more importantly, the biased datasets about Africa, prompted me to embark on an in-depth reflection on the challenges of implementing AI in Africa.
I was increasingly troubled by the human rights concerns surrounding the rollout of AI in Africa.
Just as I was delving deeper into this reflection, I encountered a young, ambitious university graduate in Bamenda who was an avid advocate for Artificial Intelligence and its potential to help him achieve his dreams.
He had applied for his dream job at a leading technology firm, excitedly uploading his resume and photo in anticipation of a seamless application process. However, he was shocked when the company’s AI-powered recruitment tool rejected his application, citing “unsuitable facial features.”
Confused and frustrated, the young man discovered that the AI system had been trained on a dataset predominantly composed of white faces.
The algorithm had learned to associate certain facial features with “employability,” unfairly penalizing him and potentially countless others with diverse backgrounds. This incident, along with others, highlights a growing concern in Africa’s rapidly evolving AI landscape: biased datasets.
As AI increasingly influences various aspects of life, from employment to education and healthcare, the need for inclusive and representative datasets has never been more pressing.
In Africa, where diversity is a hallmark, AI systems must be designed to accommodate and respect individual differences. Otherwise, we risk perpetuating existing inequalities and biases, exacerbating social and economic disparities rather than eliminating them.
If care is not taken, Africa will continue to be marginalized in the 4th industrial revolution brought about by Artificial Intelligence.
AI systems are quick to provide information about the existence of mystical orders in the Western world and their grandmasters, such as the Freemasonry, the Knights Templar, chess grandmasters, esoteric grandmasters, the Illuminati, Kabbalah, Sufism, and the Order Reaux Croix.
However, they often ignore the existence of African mystical orders like the Nantang Yoh, Kwifon, and Nwerong, along with their grandmasters. These African orders may not be as widely recognized as their Western counterparts, but they are documented and have information available online.
Besides the need for diversifying information on Africa and training unbiased datasets for the continent, African governments must be pressured to incorporate human rights impact assessments into the implementation of AI systems in Africa, where the rule of law is a significant concern.
In some African countries, governments are leveraging AI-powered surveillance systems to monitor citizens, raising concerns about privacy and freedom of expression.
For instance, in Cameroon, the government has implemented a nationwide surveillance system with the assistance of the China-based tech giant Huawei, sparking debates about the balance between security and individual rights, particularly given China’s own questionable human rights record.
Due to the ongoing conflict in Cameroon’s two English-speaking regions, minority and marginalized Anglophone communities in francophone regions and cities are subjected to undue surveillance and incrimination, often based on presumed tendencies rather than real-time crime records.
Many African countries lack robust data protection laws, leaving citizens vulnerable to data breaches and exploitation.
In South Africa, for example, the Protection of Personal Information Act (POPIA) aims to regulate data collection and processing, but its implementation is still evolving. As AI automates jobs, there’s a risk of exacerbating existing unemployment and inequality.
In Nigeria, where the youth unemployment rate is already high, the introduction of AI-powered systems could further disrupt the job market unless measures are taken to upskill and reskill workers.
AI-powered healthcare systems can perpetuate biases if trained on limited datasets.
In Africa, where healthcare resources are already strained, biased AI systems could lead to inaccurate diagnoses or inadequate treatment, further exacerbating health disparities.
As AI becomes more prevalent, there’s a risk of widening the digital divide, where those with access to AI-powered technologies reap benefits while others are left behind. In Africa, where internet penetration rates vary greatly, ensuring equitable access to AI benefits is crucial.
Foreign AI technology jobs may dominate African markets, potentially undermining local development priorities and human rights. Ensuring datasets represent Africa’s diverse populations through the development and utilization of datasets that prevent biases and inaccuracies will begin to speak to the African mind
To harness AI’s potential while protecting human rights, African governments, policymakers, and developers must prioritize inclusive datasets by ensuring datasets represent Africa’s diverse populations; develop AI systems that are transparent, explainable, and accountable; enhance laws and regulations to safeguard citizens’ personal data; prepare workers for an AI-driven job market in Africa, and finally, enable equitable access to the internet by bridging the digital divide to ensure AI benefits all.
By addressing these challenges, Africa can unlock AI’s potential to drive positive change, promote inclusivity, and empower individuals to reach their full potential. The future of AI in Africa depends on it.
*Colbert Gwain is a multiple award-winning freelance journalist, Science Fiction Technology Reporter of the Year 2024, digital migrant, and Content Creator for various media platforms in Cameroon and beyond, including print media, online websites, and local radio and television channels
