AI for Global Good
Government agencies are increasingly turning to AI to address pressing social challenges, capitalising on AI technology to analyse unprecedented quantities of data on human behaviour.
However, as the use of AI in development solutions becomes more prevalent, so does the risk of embedding bias in the design and end-use of tools. AI must be used ethically and transparently, rather than as a tool to perpetuate existing inequalities.
Our research is helping reduce embedded bias – and its consequences – in AI systems in the education sector and beyond.
Improving education attainment in Guanajuato, Mexico
Each year around 40,000 students drop out of the education system in Guanajuato, Mexico. Since 2018, the Ministry of Education in Guanajuato has been working to develop and harmonise databases on student attainment and performance to improve information about the state of the local education system.
In a pioneering pilot programme, Educational Trajectories, the Mexican government has developed an AI-based early warning system that uses this data to predict students at risk of dropping out and improve retention and graduation rates.
Itad’s research, supported by the USAID Equitable AI Challenge
In consortium with Women in Digital Transformation, PIT Policy Lab, and Athena Infonomics, we have promoted increased awareness about the importance of mainstreaming a gender perspective within the Educational Trajectories programme to mitigate AI bias.
The consortium’s work was supported by a grant from USAID’s Equitable AI Challenge, which seeks to promote innovative approaches to help decision-makers address gender biases and inequitable outcomes resulting from AI technology – ultimately fostering an equitable and inclusive digital ecosystem.
Our impact
We prevented 4% of at-risk girls from being missed
Leveraging IBM’s open-source AI Fairness toolkit, our consortium reviewed the anonymised data used to train the early alert system’s AI model. Going beyond the Secretariat of Education’s initial focus on privacy and personal data protection frameworks, our research meant we could offer actionable policy recommendations to mitigate biases and incorporate gender perspectives into AI systems.
We also facilitated a series of comprehensive workshops to strengthen the Secretariat staff’s technical expertise in the ethical, responsible and inclusive use of AI in the public sector.
As a result, our consortium discovered and prevented a critical gender bias that would have led to four in 100 at-risk girls being unidentified by the early warning system and receiving the help they needed to keep them in school.
In short, our award-winning research is helping ensure 4% more girls receive the support they need to finish their education, who may otherwise been missed due to bias identified by the program.
EQUALS in Tech Award
In 2023, our USAID Equitable Challenge project won the EQUALS in Tech Awards’ Research award, selected from more than 132 nominated initiatives from 54 countries.
The EQUALS in Tech Awards celebrates initiatives, projects, movements, organisations and institutions worldwide working to bridge the gender digital divide.
Read more: USAID Equitable Challenge project wins EQUALS in Tech Award
“We are delighted to receive this award. I hope that the resources developed through this project, benefit other development and humanitarian programmes” – Mike Klein, US Director of Itad
Facilitating global ethical AI use
The policy recommendations we provided to the Ministry of Education in Guanajuato also have actionable lessons that can be adopted by a range of diverse organisations looking to explore AI and data policies in the education, health and financial sectors.
Read a summary of our findings
Building on our engagement with Mexico, we have shared the implications of our work with government representatives in Latin America and state government leaders from Uttar Pradesh, Andhra Pradesh, Telangana and Tamil Nadu in Indi.
We encourage other regions to use lessons from the Mexico experience to replicate ethical, gender-responsive AI approaches.
Ethical AI resources
Within our support to the Educational Trajectories team, we developed an AI Ethics Guide and Checklist for AI Deployment.
We hope these resources will benefit other development and humanitarian programmes and enable them to better mitigate bias and ensure their AI projects are equitable, inclusive and transparent.
AI Ethics Guide
This guide includes an overview of what AI is, its ethical concerns, and how they can be addressed at national, sub-national and municipal levels.
It includes case study examples, a comprehensive literature review and a series of questions to prompt decision-makers to reflect on how they can ensure responsible use of AI in government systems.
Checklist for AI Deployment
In conjunction with the ethics guide, this tool is designed for policymakers and technical teams preparing to deploy – or already deploying – AI systems. It provides a starting point for building ethical AI systems, leading users through six development phases.
Further reading:
