Alexander Riabov is the Senior Communications Specialist in DAI’s Digital Frontiers. 

Disclaimer: This article was submitted to WorkwithUSAID.org as a guest blog post. The views expressed by WorkwithUSAID.org guest blog contributors do not necessarily reflect the views of the United States Agency for International Development or the United States Government.

Artificial intelligence (AI) is not neutral; instead, it involves lots of human decisions that profoundly impact how AI systems are designed and implemented. As a result, AI can work well for some and not for others. Researchers and activists have identified numerous instances of inequitable design, use, and impact of AI-enabled tools in low- and middle-income countries. Diverse collaboration through public-private partnerships can help recognize deficiencies in the technology, but encouraging these partnerships is a challenge in itself. 

USAID seeks to support early-stage experimentation, create a trusted space for meaningful exchanges, and offer direct incentives such as grants to support the private sector in emerging markets. By working with USAID, organizations can more effectively design, build, and deploy AI-enabled tools to address pressing global issues while honoring a commitment to ensuring these tools are inclusive and transparent.

In parallel, USAID and similar bilateral and multilateral donors should recognize bureaucratic processes, barriers to entry for small businesses, and divergent motivations between the public and private sectors as persistent hurdles that prevent the private sector from fully engaging in the first place. By addressing these obstacles and demonstrating the value proposition of working with governments, USAID and its donor counterparts can propel public-private partnerships to reach their full potential.

Taking on Risk So Others May Innovate

From supporting a startup that uses an algorithmic credit-scoring model to determine which smallholder farmers are likely to repay loans, to funding an AI company that specializes in supply chain forecasting, USAID’s ability to support early-stage experimentation incentivizes partners to identify and mitigate inequitable outcomes during the outset of an AI pilot. Experimentation also allows the Agency to collect evidence on what works and what doesn’t, stresses Meredith Perry, Senior Advisor for Innovation Competitions in USAID’s Innovation, Technology, and Research (ITR) Hub. 

Simultaneously, USAID can take on financial risk that some smaller companies cannot afford, says Shachee Doshi, Acting Lead for the Strategy and Research Team in USAID’s ITR Hub. Reflecting on the early stages of RappiCard Mexico, the financial technology, or fintech, joint venture of the Rappi delivery platform and Mexico’s second largest financial group Banorte, CEO José Antonio Murillo shared a similar sentiment, stating that making a difference becomes really difficult when startups have limited resources and are short on capital.

USAID also encourages others to refine their products. Highlighting a recent partnership with the Agency to develop more gender-equitable credit-scoring algorithms, Murillo notes how RappiCard expects to accelerate its learning curve on who is creditworthy. With greater insight into the company’s customer base, Murillo hopes to provide better service to a population who previously lacked access to the financial industry.

Creating a Space for Meaningful Exchanges

Committed to taking action against the unintentional consequences of AI on gender equity, USAID launched its Equitable AI Challenge—a competition designed to address gender bias and harm resulting from the inequitable design and use of AI tools. Blending competition and co-creation, the challenge encouraged participants to collaborate on solutions, identify partnerships, and strengthen their proposals—all while forming a larger community of practice.

While the challenge welcomed competitors, participants came in willing and interested to work with just about anyone, explains co-creation participant Michael Klein, U.S. Director & Vice President at Itad. Itad came into the challenge with the IBM AI Fairness 360 Toolkit and some ideas for how to use it in a development context. However, the organization lacked a really strong case study, highlights Klein. PIT Policy Lab then put forward a tangible case study that would allow Itad to use IBM’s tool in a practical setting.

What Prevents the Private Sector from Engaging in the First Place

For some private sector companies, the speed at which they work is an important consideration to ensure projects take off. Stakeholders may lose interest if collaboration is not moving at a reasonable pace, notes Murillo when reflecting on RappiCard’s engagement in working alongside USAID and a consortium of universities—two types of partners who tend to be held up by bureaucratic processes. As USAID continues to seek new private sector partners, the velocity at which partnerships are formed remains a priority for these partnerships to be sustainable.

For smaller businesses, certain contractual and bidding processes become complicated and time-consuming, shares Kate Gromova, Co-Founder of Women in Digital Transformation (WinDt Consulting). Gromova stressed that providing a different route toward financial and information resources to companies with smaller capacities or limited budgets is necessary if USAID hopes to engage organizations on a more local level.

Some innovative but smaller pilots lack certain incentives that encourage larger private sector partners to engage in AI projects. While USAID may default to a more hyper-local technical strategy, even building some AI tools as fit-for-purpose solutions that do not require growth, larger tech companies tend to favor pilots that promise scalability, highlights Josh Mandell, IBM’s Lead for Global Development. As USAID continues to seek new private sector opportunities, it is critical that the Agency and large private sector companies identify ways to integrate the strategic visions and motivations of both organizations into a partnership—ensuring a mutually beneficial relationship keen on producing cutting-edge AI solutions.

Private Sector Engagement Remains a Priority

Private sector engagement, including cooperation with smaller businesses, remains a priority for USAID as a much-needed way to address AI challenges. The Agency’s Artificial Intelligence Action Plan highlights private sector partnerships as an essential component to addressing instances of inequitable design, use, and impact of AI-enabled tools. With an opportunity to shape how technologies evolve in the developing world, a strong network of partners across public and private sectors will remain crucial to ensuring secure, open, and inclusive digital ecosystems—where all communities can benefit from emerging technologies.

For more insights on how AI is being used in development, check out this recent blog on “How USAID and Partners Can Use AI to Advance Development.”