21112 research outputs found
Sort by
Can we trust large language models to summarize food policy research papers and generate research briefs?
No full textGenerative large language models (LLMs), while widely accessible and capable of simulating policy recommendations, pose challenges in the assessment of their accuracy. Users, including policy ana lysts and decision-makers, bear the responsibility of evaluating the outcomes from these models. A significant limitation of LLMs is their potential to overlook critical, context-specific factors. For example, in formulating food policies, it is vital to consider regional climate and environmental variables that influence water and resource availability. Nonetheless, due to their reliance on word sequencing probabilities from training datasets, LLMs might propose similar policies for distinct regions. Despite these limitations, LLMs offer considerable advantages for rapid policy analysis, particularly when resources are constrained. They serve as quick, accessible, and cost-effective tools for policy research and development, requiring minimal training and infrastructure. In our study, we assessed the efficacy of LLMs in generating policy briefs by inputting an IFPRI discussion paper into three different LLM-based approaches: a standard chatbot without extra data, a Retrieval Aug mented Generation model integrating semantic search with LLM, and a custom-developed Brief Generator designed to create policy summaries from AI-analyzed paper structures. Our findings revealed that none of the LLM-generated briefs fully captured the original paper's intent, underscor ing the need for further research. Future investigations should focus on gathering more empirical data with diverse text types and volumes to better understand these outcomes.Non-PRIFPRI1CPA; Natural Resources and Resilience (NRR); Transformation Strategie
ICTforAg 2023: Cultivating inclusion
No full textICTforAg is an annual convening where agricultural stakeholders and technology experts come together to share knowledge, find solutions, and form partnerships to address challenges in agri-food systems across low- and middle-income countries. The main goal of ICTforAg is to grow communities and catalyze meaningful conversations, insights, and collaborations, increase participation of participants from the developing world, promote knowledge sharing and learning, and inspire practitioners to develop inclusive and sustainable ICT solutions. ICTforAg has a strong history since 2015 and owes its success to the contributions made by various organizations to build this community. In 2023, CGIAR and DevGlobal, in partnership with USAID Feed the Future and DAI Digital Frontiers, jointly implemented ICTforAg 2023 as a global online conference on November 7-9 with 145 speakers across 40 sessions. Out of the 2,608 individuals who registered for the event, a total of 1,778 attendees (constituting 68% of registrants) participated over the three days. In addition to the main sessions, the conference also featured the Expo, virtual exhibition space that allowed various organizations from academia, research, and the private sector to showcase their innovations interactively, the Inspire Challenge, a new Pay-for-Results program intended to increase women’s participation in digital agri-food advisory services and programs, the ICTforAg+ Satellite Events, a series of locally-led satellite events in four countries (Indonesia, Mexico, Nepal, and India), and the ICTforAg Learning Network, an online platform designed to support the collaboration amongst the global ICTforAg community of practice.Non-PRIFPRI1Natural Resources and Resilience (NRR); Transformation Strategie
AI-assisted dietary assessment and personalised “nudges” in urban youth: Feasibility and effectiveness of a mobile phone-based intervention to improve diets in Ghana
No full textDiet-related risks cause 20% of global deaths (~11 million) per year (Afshin et al., 2019). Increased urbanization in low- and middle-income countries accompanied by a nutrition transition, involving increased consumption of unhealthy, processed foods and reductions in physical activity, has exacerbated diet-related risks, leading to increases in rates of overweight and obesity (Popkin et al., 2020). Diets in urban settings are also affected by the reduced time for food preparation, convenience and affordability of unhealthy foods.
We developed an artificial-intelligence-based phone application called Food Recognition Assistance and Nudging Insights (FRANI) to tackle these problems. FRANI was previously validated against observed weighed records, the gold standard for diet assessment, and shown to be as accurate as a trained dietician undertaking a standard, multi-pass 24-hr recall (Folson et al., 2023). FRANI was developed to also provide gamified nudges and personalized feedback designed to promote healthy food choices. The objectives of this study were to measure the acceptability, adherence, and likeability of FRANI, as well as its effects on the food choices of female youth in Accra, Ghana.Non-PRIFPRI1; DCANutrition, Diets, and Health (NDH); Food and Nutrition Polic
Reverting to traditional views of gender during times of relative deprivation: An experimental study in Nepal
No full textDo individuals’ perceptions of their relative economic status affect their attitudes regarding gender roles in patriarchal societies? What role does hearing messages designed to increase support for women’s empowerment play in moderating these effects? Leveraging an original survey experiment in Nepal, we find that a prime conferring feelings of relative deprivation causes women to revert to traditional views of gender in economic decision-making; they become less supportive of women having equal control over household income, sharing house hold chores with men, and working outside the home. Women’s empowerment messaging does not attenuate these effects. Priming men to feel relatively deprived causes declines in gender equitable economic and political views, but women’s empowerment messaging nullifies these effects. The results suggest that among populations feeling relatively deprived, regressive gender norms may take hold. However, light-touch efforts to spur support for women’s empowerment may counter some reversion to traditional views of gender.Non-PRIFPRI1; 5 Strengthening Institutions and Governance; G Cross-cutting gender theme; Capacity StrengtheningPoverty, Gender, and Inclusion (PGI); Food and Nutrition Polic
Agrifood system in India - state factsheet: Bihar
No full textThis factsheet provides a brief description of the economic structure and the size of the agrifood system of Bihar, an Indian state. The agrifood system comprises primary agriculture, food processing, agrifood trade, food services, and input supplies. We have constructed a multistate social accounting matrix (SAM) for the year 2017/18 to understand the size of Bihar’s agrifood system in comparison with the rest of India.Non-PRIFPRI1; 4 Transforming Agricultural and Rural EconomiesForesight and Policy Modeling (FPM); Transformation Strategie
Nutrition-sensitive food distribution amidst inflationary shock: Evidence from a randomized intervention in Egypt
No full textWe evaluate the impacts of a traditional food distribution and a nutrition-sensitive food distribution intervention in the context of a rapidly increasing inflationary pressure in Egypt. Besides evaluating the relative and absolute impacts of these interventions on household food and nutrition security, we also examine their impacts on households’ preferences for in-kind versus cash transfers. We implement a clustered randomized control trial through which we randomly assigned communities into: (i) “nutrition-sensitive” food box, (ii) traditional “staple-heavy” food box, and (iii) control group. We find that the nutrition-sensitive food distribution cushioned falls in dietary quality and food security of targeted households relative to the control group while the impact of the traditional and staple-heavy food distribution appears to be negligible. The nutrition-sensitive food boxes increased beneficiary households’ dietary diversity by about 9 percent while also increasing energy, protein, and iron intake by 12, 13, and 19 percent, respectively. We also find that experience with the food boxes increases households’ preference for in-kind transfers, more so among households experiencing high inflation rates and among those households not covered by other food and cash transfer programs. Receiving food boxes increases preference for in-kind transfer by about 9-11 percentage points. Our findings have important implications for the debate on the efficacy of alternative interventions to support poor households as food prices rise and the relative efficacy of in-kind and cash-transfers. The lack of effectiveness of the staple-heavy food boxes suggests that the design and content of in-kind transfers are crucial when considering this policy option, including compared to cash.Non-PR2 Promoting Healthy Diets and Nutrition for all; IFPRI1; Capacity StrengtheningDevelopment Strategies and Governance (DSG); Transformation Strategie
Behavioral intelligence in plant breeding: A framework for qualitative research
No full textThis issue brief introduces a conceptual framework to describe the factors that inform farmers’ varietal uptake choices, by integrating choice behavior alongside more contextual and technical aspects of seed uptake. The framework was developed to support qualitative data analysis for generating behavioral intelligence about farmers’ decision-making about crop varieties, which may be of use to government agencies, nongovernmental organizations, and companies that operate in the seed sector.PRIFPRI1; 1 Fostering Climate-Resilient and Sustainable Food Supply; 2 Promoting Healthy Diets and Nutrition for allMarkets, Trade, and Institutions (MTI); Food and Nutrition Polic
Traders and agri-food value chain resilience: The case of maize in Myanmar
No full textMyanmar has experienced a sequence of dire crises beginning in 2019 including the unexpected closure of a principal trade route, COVID-19 lockdowns and travel restrictions, and a military coup leading to years of disruptions in the banking and transport sectors, inflation, and conflict. Yet, through these cascading shocks Myanmar’s maize sector experienced robust growth in production and exports. This paper examines the reasons underlying this apparent paradox and our findings contribute to the small but growing literatures on agri-food value chain (AVC) resilience and adaptation by traders. Strengthening the resilience of AVCs to shocks has important implications for welfare in developing countries and is increasingly drawing attention from policymakers and development partners. Using data from several sources including rare panel data sets of traders and farmers, and key informant interviews, we show that crop traders have been critical to the resilience of the maize value chain in Myanmar during this turbulent period. Maize traders performed three key functions contributing to resilience: (i) market discovery when primary trade routes were closed; (ii) overcoming transportation disruptions and bank closures to move maize from the farmgate to local and export markets; (iii) maintaining flows of credit to farmers throughout the crises in the form of selling inputs on credit and lending cash, thereby injecting much needed liquidity at times of incredible uncertainty, disruptions in the banking sector, and rising input prices.Non-PRIFPRI1; 4 Transforming Agricultural and Rural Economies; MAPSA; CRP2Development Strategies and Governance (DSG); Transformation StrategiesCGIAR Research Program on Policies, Institutions, and Markets (PIM
The contributions of Public Works in watershed rehabilitation and irrigation in Bureau for Humanitarian Assistance supported Productive Safety Net Program areas of Ethiopia to resilience and nutrition
No full textThe Ethiopian Productive Safety Net Program (PSNP), launched in 2005 and operating in eight regional states, harmonizes the delivery of donor support to vulnerable populations experiencing chronic food insecurity and shocks. PSNP investments aim to improve food consumption and nutrition, while also protecting and developing assets for sustaining stable access to food. PSNP differentiates two types of beneficiaries:
(1) Households with no able-bodied members that receive direct support in the form of cash transfers or in-kind support; and
(2) Public work (PW) beneficiaries, who receive payments in exchange for work on various types of productive investments, such as watershed rehabilitation, irrigation, road and other infrastructure.
This policy note summarizes results from an assessment of the impacts of PSNP investments supported by USAID’s Bureau for Humanitarian Assistance (BHA) in watershed rehabilitation and small-scale irrigation (SSI) on food security, nutrition and resilience in Amhara, Oromia, and Tigray regions of Ethiopia and Dire Dawa City administrative area between 2017 and 2021. The interventions assessed include area closures, SSI, other physical soil and water conservation practices, and overall watershed-level activities. Investments in SSI included construction of river-diversion schemes, lined canals, earthen canals, ponds and rainwater harvesting schemes. As watershed rehabilitation interventions take years to decades to show impacts, two time periods were used in the biophysical assessment: long-term changes, covering the period of 1984 to 2020, as well as short-term changes that specifically covered the period prior and during the USAID BHA program.
The assessment combined socioeconomic analysis (quantitative and qualitative), as well as high-resolution remote sensing and biophysical modeling in selected watersheds. Remote sensing and biophysical mod eling provided insights on environmental effects and climate resilience of BHA’s watershed rehabilitation interventions, whereas quantitative and qualitative analyses helped identify the food security, nutrition and resilience impacts of these interventions.
Of note, the assessment was undertaken during the COVID-19 pandemic and was also substantially affected by civil conflicts in the northern part of the country. This affected both the study design and data access and the implementation of the BHA-supported PSNP PW investments. The results should be seen in this context.Non-PRIFPRI1; 2 Promoting Healthy Diets and Nutrition for all; Feed the Future Innovation Laboratory for Small-Scale IrrigationNatural Resources and Resilience (NRR); Transformation Strategie
Welfare impacts of seasonal maize price fluctuations in Malawi
No full textMaSSPDevelopment Strategies and Governance (DSG); Transformation Strategie