On the 12 November 2020, I had the rare privilege of being part of a panel discussion on the “opportunities and challenges for Responsible AI in Afric
On the 12 November 2020, I had the rare privilege of being part of a panel discussion on the “opportunities and challenges for Responsible AI in Africa” as part of the Responsible AI Forum Preview organized by the Institute for Ethics in Artificial Intelligence (IEAI) in cooperation with the Global AI Ethics Consortium (GAEC) and the Responsible AI Network-Africa (RAIN Africa). As panelists, we unanimously agreed that there is a lack of access to quality, unbiased data for implementing AI solutions across Africa. However, from my interactions with startups in the agriculture sector, I am convinced some amount of data exists scattered with different organisations. The challenge is that no framework allows consolidation and sharing in a way that benefits the data service provider, contributor, and user. In part one of this article, I will attempt to discuss several data initiatives that will assist in dealing with this challenge in Ghana.
Data commons are initiatives in which data is shared as a common resource among individuals and organisations who jointly agree on a data governance framework for accessing it. To create a data commons, three main stakeholders are required. A data commons service provider, data contributor, and data user. It is also necessary to have data contributors and data access agreements that provide guidelines for managing and accessing the data respectively. Data commons makes data available to stakeholder which helps in advancing a field quickly. In areas such as medicine, data commons are essential because a critical mass of data is required for evidence. An interesting data commons is Dataverse. Dataverse allows the sharing, preservation, citation, exploration, and analysis of research data. Data commons also benefit from network effects through “data peering” where data commons service providers agree to allows each other’s data user access at no cost.
Data Exchanges and Markets
With the emergence of Big Data, data is has become is a valuable commodity. Data exchanges and markets are, therefore, platforms that treat data as an economic good. Data marketplaces sell different types of data from several sources. They offer incentives including cash to promote data sharing. Pockets of data sitting on devices of different data providers do not a lot of value. Aggregated data provides the incentive for providers to share data. Usually, the data shared by data providers with the data marketplace is aggregated and anonymized and presented to data consumers as data-based services. The data consumers pay for these services which enable the data marketplace to provide incentives for the data providers to share data. This model provides a win-win situation for organisations to monetise their data and others to use the data saving them time and the cost of collecting the data from scratch. A good example of data exchange is the Copenhagen- Hitachi City Data Exchange which is a public/private partnership between the Copenhagen Municipality and Hitachi. This collaboration involves purchasing, selling, and sharing data types between citizens, public institutions, and private companies in the city.
Open Data Platforms
Open data platforms are curated sets of open datasets. Ghana Open Data Initiative (GODI) is led by the National Information Technology Agency (NITA) under the Ministry of Communications of Ghana. The project was started in 2012 with support from the Web Foundation. While this initiative is good, granting access to public sector data should not be the end. Strategies must be put in place to encourage usage and improvement of the data that is collected. The Ghana Open Data portal can be improved in several ways. To attract users and bring the users closer to the public sector information, a data hackathon was organised in 2019. In addition to this, the portal must provide quality and updated data to retain old users and attract new ones. To make the portal user-friendly, the datasets must also be visualised. Finally, the datasets must also be made accessible in search engines.
Big data is driving innovation in science, technology, and business. Data is a critical resource for solving some of the challenges we face in agriculture, health, education, etc. Therefore, the ability to gather, analyze, and use data is rapidly becoming the focus of governments around the world. More important we must find innovative means to share the data that individuals, organisations, and the government has. While there is the need to collect data across all sectors of the economy, I believe that with the right data governance approach, we can benefit immensely from the little data we already have.
Author: Kuuku Sam is Advisor, Artificial Intelligence for Sustainable Development at GIZ and Executive Member, Institute of ICT Professionals, Ghana.
For comments, contact Kuuku on 0274 333 510 or at firstname.lastname@example.org