Discover the latest trends

In 2021, the Centre for the Fourth Industrial Revolution (C4IR) conducted two pilot projects to test the AI Procurement in a Box Guidelines, a guidebook designed to facilitate the adoption of AI in the public sector through government procurement, along with a set of complementary tools that illustrate the growing global consensus on the responsible use of AI technologies. 

The first pilot was with the Metrô de São Paulo and the second was with the Hospital das Clínicas at the University of São Paulo.  These pilots demonstrated the need for guidance on how to ethically and safely purchase artificial intelligence and machine learning (AI/ML) tools, and also identified challenges in using these technologies in the public sector. 

Hospital das Clínicas (HC) of the University of São Paulo’s Use Case

This pilot project examined the hospital's technical and data access capabilities, with a focus on the ethical use of AI in healthcare, a field under close public scrutiny. 

HC aimed to integrate its 60+ information technology (IT) systems into a cohesive system, allowing it to effectively utilize the data it collects in other areas such as teaching, research, and innovation.

Significant organizational changes were implemented, a new platform was developed to improve privacy and data management, and a culture of engagement with all people involved in the data process was promoted.

Metrô de São Paulo’s Use Case

The pilot conducted by the Metrô de São Paulo, a major subway system with over 5 million daily passengers, revealed the difficulties of implementing an innovative procurement process for the first time and the use of an algorithm impact assessment tool to identify and mitigate potential risks in deploying a predictive maintenance AI tool.

The goal of the Metrô pilot was to purchase an AI-based predictive maintenance system for continuous online monitoring of the subway's rail tracks. High-definition cameras on trains and sensors would allow for the use of deep learning and automated image processing to assist in decision-making by Metrô's engineering team.

As an extension to this original project, Rafael Carvalho de Fassio, Artificial Intelligence and Machine Learning Fellow seconded from the State Attorney-General's Office in São Paulo, Brazil, and Clara Clemente Langevin, Project Lead for Artificial Intelligence and Machine Learning at Centre for the Fourth Industrial Revolution, Brazil, led the creation of the whitepaper "Unpacking AI Procurement in a Box: Insights from Implementation." This whitepaper discusses the challenges that need to be addressed in order to make the AI Procurement in a Box toolkit accessible to countries in the Global South. It also includes a framework for monitoring and evaluating future pilot projects, allowing for comparisons of the success of implementation and the dissemination of best practices in the public sector across countries and regions.

Let's take a look at some of the insights from the report:

📌 “Due to the loss of mobility during the COVID-19 pandemic, many countries rushed to digitalize government services” 

📌 “In Latin America alone, the number of citizens accessing the internet for government transactions jumped from 21% to 39% during the pandemic”

📌 “Government reliance on emerging technologies such as AI is set to grow exponentially, so it is imperative to prepare institutions around the globe through best practices such as those associated with AI Procurement in a Box to best use it”

📌 “With the resurgence of new literature and new pilots, the Forum and its Centre for the Fourth Industrial Revolution Network have identified new topics that need to be addressed in future pilot” 

📌 “As AI technologies evolve, it is also necessary for governance structures to encompass new issues, as well as ensure the inclusive implementation of AI/ML worldwide” 

📌 “The experiences of piloting the project in Brazil can support nations in the Global South as they address the challenges in adopting public sector AI” 

📌 “The pilots [Hospital das Clínicas (HC) - Metrô de São Paulo] and their outcomes add to the overall AI Procurement in a Box project by exploring elements to facilitate AI adoption in the Global South and incentivize debates within the public sector on how to procure AI responsibly and efficiently”

📌 “As more countries consider adopting AI Procurement in a Box, more elements on how to best to do so and build upon these best practices will come to the forefront. This project will therefore continue to evolve – just like the technology it aims to demystify – to ensure that different countries with different contexts can effectively adopt it, even at different government levels” 

To learn more about the valuable insights from the whitepaper, we invite you to read "Unpacking AI Procurement in a Box: Insights from Implementation" Simply click the link to access the report.