Author : Aleksandra Gadzala
Type of publication : Issue Brief
Date of publication : November 2018
Artificial intelligence (AI), which enables machines to exhibit human-like cognition, is unleashing the next wave of digital disruption. Global investment in AI skyrocketed to somewhere between $20 billion and $30 billion in 2016. While many uses of AI are still in the experimental phase, commercial applications are already surfacing in a variety of sectors: AI systems filter emails, recommend items for purchase, provide legal advice, and drive cars.
The success of mobile technologies (tech) across Africa is prompting speculation among tech investors about whether AI applications will also take root in African nations. Mobile technologies, after all, have permitted African nations to dramatically increase their communication capabilities while leapfrogging the need for old-fashioned infrastructure.
An Uneasy Environment for AI
At its most basic, artificial intelligence uses algorithmic techniques loosely modelled on the human brain to en- able machines to discover patterns, generate insights from the data to which they are exposed, and then apply those lessons learned to future decision making and predictions. AI is also now being used in more complex applications including the analysis of large genome sets in an effort to prevent diseases, and the mapping of human mobility patterns to predict and control humanitarian crises.
To perform such functions AI depends on robust digital foundations, which include the availability of large volumes of data—usually referred to as “big data.” Machines can analyse this data to learn, make connections, and arrive at decisions. But AI also relies on significant know-how among its human adopters: industry leaders must know how to successfully implement AI into their operations, and consumers must be comfortable with its use (and that includes a reasonable assurance of data privacy). With the exceptions of Kenya, South Africa, Nigeria, Ghana, and Ethiopia — where these factors are rapidly coming together on the back of other enabling factors — most African countries currently struggle to meet any or all of these requirements.
Globally, Africa has the lowest average level of statistical capacity. Only half of African countries have carried out more than two comparable household surveys in the past ten years and statistical capacity has, over the last fifteen years, declined more in Africa than in any other region of the world. The weaknesses of the data are expressed in the instability of even headline statistics like economic growth and population size.
Lacking or faulty data severely limits the efficacy of AI systems. Discrepancies between on-the-ground realities and data input into AI systems may cause systems to learn incorrectly, yielding erroneous outputs. Skewed input data additionally opens the door for the reproduction and even amplification of human biases and discrimination—an issue that is especially sensitive in countries like Ethiopia and Kenya where long-standing ethnic tensions continue to inform politics and business.
But AI also relies on significant know-how among its human adopters: industry leaders must know how to successfully implement AI into their operations, and consumers must be comfortable with its use (and that includes a reasonable assurance of data privacy)
The African Union’s (AU) Convention on Cybersecurity and Data Protection encourages African governments to recognize the importance of data security and provides a framework for leaders to integrate into their respective legislations. Adopted by the AU in 2014, the Convention has yet to take effect as only ten out of the fifty-four AU member states have ratified it. Data privacy groups worry that some governments may have a vested interest in obstructing such regulations in order to access citizen data — or that, if introduced, regulations may curb free speech.
Concerns over data privacy are met with additional anxieties over automation and potential job losses resulting from the adoption of artificial intelligence solutions. Without a clear understanding of the potential advantages of AI solutions and a workforce able to onboard and take advantage of AI solutions, demand for AI in most African countries is likely to remain low.
The skills base of the continent’s workforce is lower than that of any other global region as indicated by the World Economic Forum’s (WEF) Human Capital Capacity Index, which reflects the percentage of a region’s workforce that has attained tertiary, secondary, and primary education as well as the percentage that has literacy and numeracy skills. Ethiopia is one of the lowest performers on the Index, fourth from the bottom (above only Senegal, Mauritania, and Yemen). Nigeria also ranks in the bottom twenty countries of the WEF index, at 114th out of 130 countries. With an estimated 10.5 million children out of school, Nigeria’s unenrolled rate is the highest in the world,19 and its primary and secondary school systems are largely failing.
Without significant improvements, neither Ethiopia nor Nigeria can hope to produce an AI-ready workforce, which must be well versed in advanced digital skills and data science, and also in complementary disciplines like economics and psychology that become more important as more jobs become automated.
Even in “developed” markets, few firms have so far deployed AI at scale, as they remain uncertain of either the business case or the likely return on investment. In African countries where the data ecosystem and infrastructure are wanting, and the workforce is not yet equipped with the skills necessary to adopt and advance AI solutions, the case for the widespread adoption of the technology is often even less clear.
Data privacy groups worry that some governments may have a vested interest in obstructing such regulations in order to access citizen data — or that, if introduced, regulations may curb free speech
Where AI Succeeds
Significant hurdles notwithstanding, AI solutions are being successfully deployed at scale in some African countries and especially in Kenya, Nigeria, Ghana, Ethiopia, and South Africa. Most solutions currently target the financial services, agriculture, and healthcare sectors.
South Africa leads the continent in AI adoption with a robust ecosystem that includes numerous technology hubs, research groups, and forums like the AI Summit. There are an estimated one-hundred-plus companies in South Africa that are either integrating AI solutions into their existing operations or that are developing new solutions using AI.
In Nigeria, a chatbot called Kudi AI is integrated into Facebook’s Messenger app, and facilitates mobile banking and payment services to users who may not have access to, or may be unfamiliar with, browser-based online banking but are comfortable with text-based messaging. In like manner, MomConnect, a chatbot initiated by South Africa’s National Department of Health, connects an estimated 1.8 million expectant mothers with pre-and post-natal services. Registered women are able to “chat” with the app and receive healthcare advice relevant to their pregnancy.
AI holds promise in countries where governments have made technology a national priority and are taking concerted measures to stimulate innovation and to improve data protection, research, and development. Not surprisingly, these are the nations in which AI technologies are beginning to pay off.
A growing number of African governments are beginning to realise that they cannot on their own fulfil their development goals: commercial technology solutions will play a bigger role in fixing issues previously entrusted to government bureaucrats and aid agencies. Like Kenya, Ghana is one of the few African countries with an open data initiative. In order for technological innovations to be successful in the long term, progress and innovation in government policies is also necessary.
Among the reasons behind Google’s decision to locate its AI research lab in Ghana is the country’s “strong ecosystem of local universities” and expanding network of technology hubs. Ghana has twenty-four tech hubs—the most in sub-Saharan Africa after South Africa (59), Nigeria (55), and Kenya (30)—ten public universities, and several private institutions, many of which maintain partnerships with universities around the world.
In Nigeria, a chatbot called Kudi AI is integrated into Facebook’s Messenger app, and facilitates mobile banking and payment services to users who may not have access to, or may be unfamiliar with, browser-based online banking but are comfortable with text-based messaging
Universities and tech hubs are key channels through which African countries can acquire, domesticate, and diffuse new technologies in the economy. Even more importantly, they are critical for identifying emerging technologies that can serve as a platform for producing new products and services to address local challenges.
Connecting universities with tech hubs promotes innovation in AI and other disruptive technologies by bringing together teaching, research, product development, and commercialization—functions that are often kept separate. The value in this approach is that it creates full value chains for specific AI solutions and connects key stakeholders through continuous interaction.
Les Wathinotes sont soit des résumés de publications sélectionnées par WATHI, conformes aux résumés originaux, soit des versions modifiées des résumés originaux, soit des extraits choisis par WATHI compte tenu de leur pertinence par rapport au thème du Débat. Lorsque les publications et leurs résumés ne sont disponibles qu’en français ou en anglais, WATHI se charge de la traduction des extraits choisis dans l’autre langue. Toutes les Wathinotes renvoient aux publications originales et intégrales qui ne sont pas hébergées par le site de WATHI, et sont destinées à promouvoir la lecture de ces documents, fruit du travail de recherche d’universitaires et d’experts.
The Wathinotes are either original abstracts of publications selected by WATHI, modified original summaries or publication quotes selected for their relevance for the theme of the Debate. When publications and abstracts are only available either in French or in English, the translation is done by WATHI. All the Wathinotes link to the original and integral publications that are not hosted on the WATHI website. WATHI participates to the promotion of these documents that have been written by university professors and experts.