United Nations Children’s Fund (UNICEF)
United Nations Children’s Fund (UNICEF)
Description of Activities on AI
Project 1: Using human mobility data to create risk maps for the spread of diseases
Creating risk maps for the spread of diseases. This is to help minimise the spread of diseases especially among children.
Project 3: Together with ESA and WFP, mapping crops in Malawi using drone imagery and AI
Mapping crops in Malawi to improve forecasting agricultural supplies, improve collecting crop production statistics, facilitating crop rotation records, mapping soil productivity, identification of factors influencing crop stress, assessment of crop damage due to storms and drought, and monitoring farming activity.
Project 5: Assessing biases in AI for satellite imagery classification
AI algorithms to identify objects in satellite imagery are trained using ground based samples with labels. Biases in the AI arises when such ground based samples are geographically or thematically imbalanced. The project investigate the biases in AI algorithms in popular datasets in terms of accuracy and analyze the potential drivers of such biases using various socio-economic covariants.
Project 6: Improving the quality of population data using AI and machine learning
High resolution population data is gaining its importance in data analysis for humanitarian operations. There are several available high resolution population data produced using AI and satellite imagery. This project investigates the errors and biases in those high resolution grid population data sets and develop a method to produce an improved population data set using AI and machine learning algorithms.
Project 7: Air quality prediction using AI and machine learning
> 500,000 of children under 5 years old died from air pollution related causes in 2016, and millions more suffer from respiratory diseases that affect their development, Children are uniquely vulnerable to air pollution – due both to their physiology, and to the type and degree of their
Exposure. Before and during COVID-19 we see dramatic changes in the air pollution levels in different places which might significantly impact on children’s health. In this project, AI model to make predictions of air pollution level using remote sensing data and available ground observations are developed to provide accurate air quality monitoring where the ground observations are rarely available.
Project 8: Mapping infrastructures using AI and satellite imagery
Information on the locations of infrastructures such as schools and health centers has a great importance in making plans for field operations. School locations are being mapped in this project using AI and high resolution imagery and ground observation gathered from ministry of educations from partner countries and open source data platform such as open street map.
Project 9: Covid-19 impact analysis using mobile data and AI/machine learning algorithms
Impacts of Covid-19 on how people behave are identified using mobile data and machine learning algorithms. The patterns are analyzed by the different socio- economic groups to gather insights on how people change behavior based on their socio economic status which can provide critical information for the policy making process to improve society’s resilience to the pandemic.
Project 10: Collaborating with academia to create new techniques that can help reach the most vulnerable children
Ensuring that every child can benefit from the potential of AI by providing young people with the information, skills and services they need to shape the future they want.
Project 11: Responsible Data for Children
It seeks to build awareness regarding the need for special attention to data issues affecting children—especially in this age of changing technology, data linkage and AI; and to engage with governments, communities, and development actors to put the best interests of children and a child rights approach at the centre of our data activities. Drawing upon field-based research and established good practice, RD4C aims to highlight and support best practice in our work; identify challenges and develop practical tools to assist practitioners in evaluating and addressing them; and encourage a broader discussion on actionable principles, insights, and approaches for responsible data management.
Project 12: Common messaging on the use of biometrics (Legal Identity Agenda)
UNICEF is currently working with UN partners across the system through a biometrics working group established under the UN Legal Identity Agenda (co-chaired by UNICEF, UNDESA and UNDP). The intent is to work towards a common position/ messaging on the use of biometrics across the broad areas of engagement (legal ID, security, and functional identification/ sectoral systems) – with a view towards developing shared principles.
Project 13: Guidance on appropriate us of Biometrics
The guidance outlines key questions that should be asked before determining whether the use of biometric technology is appropriate for a proposed program. Covers areas such as the program intent, target audience, data privacy and protection, community sensitivities, and technology performance. Reference: Nicola Richards (2019) Guidance on assessing the value of including biometric technologies in UNICEF-supported programmes, Data and Analytics, UNICEF, NY.
Project 14: Publications and Guidance Related to AI/Big Data/Children
Biometrics and Children: A literature review of current technologies – prepared by UNICEF and the World Bank (Forthcoming), Berman, Gabrielle; Carter, Karen; Garcia Herranz, Manuel; Sekara, Vedran (2020). Digital Contact Tracing and Surveillance During COVID-19. General and child-specific ethical issues, Innocenti Working Papers no. 2020-01, UNICEF Office of Research – Innocenti, Florence Berman, Gabrielle; de la Rosa, Sara; Accone, Tanya (2018). Ethical Considerations When Using Geospatial Technologies for Evidence Generation, Innocenti Discussion Papers no. 2018-02, UNICEF Office of Research – Innocenti, Florence Berman, Gabrielle; Albright, Kerry (2017). Children and the Data Cycle:Rights and Ethics in a Big Data World, Innocenti Working Papers no. 2017-05, UNICEF Office of Research – Innocenti, Florence Berman, Gabrielle; Powell, James; Garcia Herranz, Manuel (2018). Ethical Considerations When Using Social Media for Evidence Generation, Innocenti Discussion Papers no. 2018-01, UNICEF Office of Research – Innocenti, Florence
Project 15: UNICEF Policy Guidance on AI for Children
A guide for creating and implementing AI policies and systems that protect children’s rights and brings the attention of the public and private sectors to how AI systems impact on children. To develop the guidance over 200 experts have been consulted in 5 regions, and almost 250 children have been consulted on AI issues.
Project 16: Enhancement of flood early warning system in Malawi
UNICEF is increasingly use geospatial-mapping techniques to predict the flow patterns of floods and the new trajectories of storm zones to help Governments make data backed decisions to re-locate vulnerable populations to higher ground, re-locate schools, health facilities and key infrastructure in the pathways of repeated storms to help keep them out of harms way and operational for the communities.
Project 17: Artificial Intelligent solution for skills and job matching
Bashar soft (Vendor) (Wuzzuf/Forasna) has been selected by UNICEF private sector in Egypt Office as the most suitable solution in Egypt and proposal has been received from the company. UNICEF, MOYS (Ministry of Youth and Sports) and basharsoft will work together on the different activities, responsibilities, duties and liabilities on each partner. Bashar Soft will use their strong position in the Egyptian market and their wide database of job seekers and companies to promote the Meshwary program (a national skills development and career guidance programme that has been implemented since 2008 by the Ministry of Youth and Sports (MoYS) with the support of UNICEF) and its values and objectives to relevant job seekers and all hiring companies to increase the awareness about the amount of impact it has on raising employability levels
Project 18: EQUIST
EQUIST (EQUItable Impact Sensitive Tool) is a powerful web-based analytical platform designed to help decision-makers and programme managers develop strategies to address health system barriers and bottlenecks for achieving health and nutrition outcomes for the most vulnerable children and women.
Project 19: Targeting HIV interventions for adolescents and young people
Combining traditional and big data sources through novel analytics to prioritize sub-national geographies and segment population 15-24 years, to end HIV among young people.
Project 20: UNICEF’s Good Governance of Children’s Data
Through the initiative a number of papers on emerging AI and data related issues are being written, such as on child rights and data protection by design, state surveillance and responsible group data. for children.
Project 21: Automating U-Report helpline responses to increase scale of impact and reduce resource requirements
Increase ability of young people to access accurate and timely information on topics like SRH, MHM, using their own natural language.
Project 22: Classify rumors and misinformation on COVID-19 from multiple sources, and provide debunking information to chatbot users
Increase UNICEF’s ability to track COVID-19 related misinformation and develop more accurate solutions for providing debunking facts to chatbot users.