STEP 1: PREPARING FOR CO-CREATION
Glossary of Actors
Who might be involved in the co-creation of an AI product?
Host country governments
Who: Federal ministries, sectoral governing bodies/institutions, provincial and municipal/city governments, in-country ministries, local owners of data (e.g. governmental admin data), politicians, lobbyists
Considerations: Host country governments can provide official statistics and support in the alignment of AI/ML projects with sectoral and/or digital priorities. Their engagement can help verify policies and standards and identify opportunities for scaling. Political and economic cycles are important to be aware of, with proximity of elections and budget announcements - plus generally sensitivity to public reputation and a greater level of risk aversion - likely to influence actors’ capacity and/or willingness to engage in a project.
Civil society
Who: Implementing partners, community groups, academia, other active NGOs, general public, end-users (the people who experience the problem we are trying to solve), journalists, people whose data is used as inputs to the model, people in the community who might not use / benefit directly from the product but may be indirectly impacted, universities or other research institutions
Considerations: Civil society actors can help build knowledge and trust, and can ensure an accurate representation of affected stakeholder segments. Civil society can also aid with the dissemination of products and programmes as well as with building and reinforcing community acceptance. However, challenges including discrepancies in contextual and technical knowledge must be kept in mind when engaging with different civil society actors.
“As is true of many small companies, we can’t necessarily afford a highly extensive set of protocols, so we tend to be dependent on whoever is paying for the work”
“The ethical/responsible use of technology and AI ultimately comes down to how we enable human rights better through technology”
Private sector
Who: Local technology firms, startups, telecoms, social media companies, financial institutions, other data-holders/sectoral actors, legal experts.
Considerations: The private sector can help with data collaboration or licensing, as well as with providing technical expertise or support with capacity building initiatives. However, it is important to acknowledge the interest in profit-generation that private sector companies might have which might prevail over certain ethical considerations. Additionally, there may be limited familiarity among these actors with the local context and/or target end user groups of the AI product, potentially requiring additional engagement or facilitation to ensure the implications of outcomes produced by a model are thoroughly understood.
Donor community
Who: Funding agencies, local/international/multilateral donors.
Considerations: The donor community can help convene and facilitate partnerships and provide guidance on best practices. This can help capture and disseminate lessons learned with the broader community. However, donor agencies are often located in the Global North and might not possess an accurate understanding and knowledge of local contexts.
“Using emerging tech ‘responsibly’ means not being driven by the fact that you have the opportunity to use technology, but that you are using it with a clear understanding of what you’re trying to solve by doing it and being intentional about its use.”
Technical actors
Who: technologists, tech NGOs, local tech service providers, Local tech community, university graduates with technical skills, developers, data scientists, IP owners, human rights experts
Considerations: Involving technologists in the co-creation process is the first step in having an efficient and valuable AI product. To do this, one must create a common language or taxonomy that can be understood in the tech sphere. User cases and guidance on how to implement specific guidelines are also useful. Technologists must understand how to test specific propositions, and how to translate policy into tech as the lack of specific guidelines often impedes motivation. One way to go around this is to provide direct funding and support to tech NGOs to increase their incentives, and bridge the gap between tech experts and the policy and business worlds. Software development kits (SDKs) are useful in doing this, as well as framing policies using a Business Process Model and Notation (BPMN) model.