The use of artificial intelligence (AI) in international development is evolving from theoretical discussions to practical applications. At the upcoming 2024 Australasian AID Conference, one session focused on the exploration of AI tools that are being developed to improve development policy and programming. This has significant implications for how aid organizations analyze challenges and make decisions.
Two organizations, Dragonfly Thinking and the Development Intelligence Lab, showcased how structured AI applications can enhance human analysis while mitigating common issues associated with AI implementation. Miranda Forsyth, co-founder of Dragonfly Thinking and a professor at ANU, shared her transition from AI skeptic to innovator. Despite her initial reservations, Forsyth collaborated with Professor Anthea Roberts to create an AI platform aimed at tackling complex development issues.
The core of their approach is the “Triple R Framework,” which considers risks, rewards, and resilience, unlike more traditional methods that may overlook system dynamics over time. This framework captures risks at the crossroads of threats and vulnerabilities, while rewards connect opportunities and capabilities.
The resilience aspect of the framework includes three capacities: absorptive for immediate responses, adaptive for real-time strategic adjustments, and transformative for broader systemic changes. Dragonfly’s AI tool integrates over large language models, guiding users through structured analysis and incorporating diverse perspectives to provide a holistic view.
Forsyth demonstrated the platform’s ability to analyze sorcery accusation-related violence in Papua New Guinea, illustrating how it can help identify social media’s role in propagating or curbing violence. Meanwhile, the Development Intelligence Lab has been testing AI in partnership with human analysts to improve policy evaluation. CEO Bridi Rice noted that in a recent exercise, a team without a development economics background was able to produce valuable insights in just two hours using the Dragonfly platform.
The emphasis of these organizations is not solely on efficiency improvements, but rather on the enhancement of analytical depth and decision-making quality. They recognize the challenges of AI bias, particularly highlighting the need to include local perspectives in discussions that often favor Western knowledge frameworks.
With projects like the “Southeast Asia Pulse Check,” which collects insights from numerous experts across various countries, there is potential for AI to amplify local voices within the development dialogue. The technology’s capacity for analyzing multiple viewpoints can help align development efforts more closely with the realities faced by local populations.
Looking ahead, both organizations plan initiatives for 2025, including the formation of an “AI in development” discussion group and partnerships aimed at addressing policy challenges with AI technologies. They are committed to integrating diverse knowledge sources, including indigenous perspectives, into their frameworks.
However, pressing challenges remain, such as managing AI’s energy consumption and ensuring meaningful incorporation of indigenous knowledge. Both Forsyth and Rice stress that AI should complement, not replace, extensive local engagement in program design. Maintaining human oversight as AI tools advance is crucial, as these technologies aim to enhance—not supplant—human judgment.
The development community must also confront issues of access and equity, particularly ensuring that smaller organizations and those in developing regions can leverage these innovations despite potential resource constraints.
In conclusion, the collaboration between Dragonfly Thinking and the Development Intelligence Lab exemplifies how thoughtful integration of AI into development practices can empower practitioners to tackle complex global challenges more effectively. The advancements in this arena suggest a promising future where technology can bridge the gap between local realities and development goals, fostering more informed and equitable decision-making.
This development has the potential to not only improve the efficacy of aid organizations but also ensure that the voices of those most affected are heard and incorporated into crucial policy-making processes.
Leave a comment