As AI continues to grow, it is being utilised more and more in Kenya. AI has the potential of boosting the country’s economy and solving prevalent problems across various sectors. Its application can transform the delivery of government services, improve the healthcare system, increase food supply, and enhance education, among others. According to research conducted by CIPIT on the utilisation of AI in Africa, Artificial Intelligence has 49 applications in Kenya. These applications cut across 15 sectors which include agriculture, health, finance, education, businesses, and security.

Key industries utilising AI in Kenya

Kenya’s agricultural sector has several AI applications designed to meet various needs. These include the Third Eye project which uses drones to monitor soil needs and identify pests and diseases in crops. It detects minute changes in crops that are not easily discerned by humans and enables early detection which saves crops. Other notable applications are Nuru22 and Eska. These applications rely on AI in detecting pests and diseases in crops.

AI is heavily relied upon in fintech (digital lending apps) in determining whether or not to grant loans to borrowers.  AI through machine learning, in this case, is used to determine a borrower’s likelihood of default through the assessment of their behavioural data and information obtained from their device including contacts, SMSs, SD card content, gallery, and apps downloaded in the device. The information is used to determine a borrower’s credit score and determine suitable loans for them. Fintech relies on this information because they don’t obtain security from borrowers compared to traditional financing institutions such as banks. Their business model enables them to provide loan facilities to the unbanked population which may lack security to access loans from banks.

AI is being used to address challenges faced in the medical sector. It is being used in rural areas for example, in the detection of diseases such as cervical cancer in women. Tambua Health is also another application that is employed by medical personnel in the diagnosis and treatment of cardiopulmonary diseases. It relies on machine learning to analyse the sound of a patient’s lungs and heart to detect cardiopulmonary diseases.

Risks posed by the growing adoption of AI in Kenya

AI poses the risk of rendering many people jobless due to automation. This is because most routine tasks will be taken up by automation and the people performing these tasks will be left jobless. According to the Future of Jobs Report by the World Economic Forum, 52% of jobs in Kenya are susceptible to automation.

The adoption of AI also provides governments in Africa with an avenue to surveil citizens. This is mainly enabled through facial recognition tools that monitor the activities of citizens, profile them, and locate them.

AI promises significant benefits to the country’s economy, however, these benefits stand a risk of not being fully realised due to the gender inequality faced by women in terms of AI use. Gender inequality by AI is perpetuated in several forms. These include bias against women caused by AI algorithms, the low number of women in the AI ecosystem that is academia and AI workforce, and the low number of women accessing the internet and owning mobile phones coupled with the low levels of digital literacy. Bias against women is caused by several factors such as training AI algorithms with datasets that are inaccurate, inadequate, under-representative, and historical. It is also caused by the transmission of bias in the systems by developers, which in most cases is unintentional. Their heavy reliance on personal data raises concerns about privacy and data protection.

AI is relied upon for automated content moderation in online spaces where it is applied to detect, curate, and remove content. Its use involves the training of algorithms through machine learning models which enable them to take automated actions and make decisions regarding content in online spaces. It is employed in content moderation mainly to target hate speech, disinformation, and content that affect democratic processes. As much as it can be effective in regulating content, it also carries the risk of undermining human rights (such as freedom of expression and access to information) by taking down relevant content and leaving unwanted content online.

How to create a framework for AI in Kenya

According to The State of AI in Africa Report by CIPIT, developing a national AI strategy is one of the key steps in regulation. A national AI strategy will guide the development of skills, technology, ecosystem, infrastructure, and governance mechanisms to ensure its sustainability and competitiveness. The AI strategy should promote a human-centric AI approach where people’s well-being is a priority. It should also facilitate a multi-stakeholder dialogue on the deployment of responsible AI for the benefit of society and to inform related policy discussions.

Responsible AI should also extend to responsible data systems. Data governance structures must therefore consider elements of responsible AI, particularly in the development stages of AI systems. Mechanisms must also be put in place to monitor accountability and compliance in relation to ethical and responsible AI.