Why Kenya’s AI Strategy Is Bigger Than Technology
To bring these issues into perspective, it is useful to examine a few. The Strategy recognises that Kenya lacks a comprehensive AI-specific policy framework, an absence that could hinder research, innovation, investment, and adoption. A relevant example is the copyright challenge faced by Masakhane, an organisation championing Natural Language Processing (NLP) research for African languages. Masakhane had been using the JW300 dataset, a multilingual corpus of over 300 African languages derived from Biblical texts produced by the Jehovah’s Witnesses, with an average of 100,000 parallel sentences per language. However, after seeking legal advice from the Centre for Intellectual Property and Information Technology Law (CIPIT) in Nairobi, Masakhane learned that the dataset’s copyright notice did not permit text and data mining (TDM). A formal request for permission was declined, prompting Masakhane to discontinue use of the dataset and redirect efforts toward building their own datasets through ethical, community-driven initiatives. This case highlights the urgent need for clear legal and ethical frameworks that support local AI research and innovation.
