AI for Good: How Google is building AI solutions to global problems in Africa

In June 2018, Google announced plans to launch an artificial intelligence centre in Africa. And, the company finally did a few months later, in Accra, Ghana. 

For Africa and managed by Africans, the AI centre was designed to demonstrate its commitment to solving global crises that are prevalent on the continent through developing innovative ideas that are driven by data.

Traditionally, Google focuses on fundamental research in computer science and related fields, with teams located all over the world. The tech giant extended its research presence to the African continent by opening an AI research centre in Accra, Ghana, in 2018 and expanded its frontiers to Nairobi, Kenya, in 2022.

The Google Research Africa teams are focused on developing AI solutions that are relevant to the African continent.

Read as well: Google Officially Launches AI Centre in Accra, Ghana

Last week, the team shared details of some of its AI projects with journalists from South Africa, Nigeria and Kenya at its Research Africa office in Accra, Ghana this week.  The team demonstrated AI solutions to challenges of flooding, food security, identity management and natural disasters.

AI for Africa, made in Africa

Speaking at the event, Yossi Matias, VP, of Engineering & Research, at Google said that Google’s investment in Africa is a reflection of the continent’s growing importance in the global technology landscape. Africa is home to an increasingly tech-savvy population and the continent is rapidly becoming a major market for digital products and services.

He shared his belief that AI has the potential to be a transformative technology for the continent.

According to him, the typical flow dictates that innovative solutions to problems are developed for other climes and shipped to Africa. However, the Google Research team in Africa is focused on developing Al solutions that will solve problems that are peculiar to the African continent and that have the potential for global relevance. 

Yossi told the audience that Google’s investment in research in Africa is a reflection of the continent’s growing importance in the global technology landscape. 

According to him, since 2018, Google has through its academy trained thousands of tech talents, some of which have been retained by Google while others have founded startups or joined other companies.

Co-leaders of the Google Research Africa team, Jason Hickey (PhD) and Aisha Walcott (PhD) pointed out that of many ideas in the pipeline, the team has selected to build AI solutions to solve the problems of flooding, food security, weather and geolocation on the continent. And, this is just a starting point.

AI for weather forecasting

According to the World Meteorological Organisation (WMO) over the past 50 years, 22% of deaths and 67% of economic losses caused by natural disasters were the result of extreme precipitation events.

“ You also saw the tremendous potential that having accurate weather forecasts for Africa can have in terms of agriculture and other industries that rely on weather for their decision making, not only for industry but also down to the individual,” Walcott said.

Google’s weather forecasting tool has been built on machine learning models that precisely estimate current precipitation conditions. According to the project lead, Emmanuel Brempong, the model has been used to provide early warnings, especially in Africa where rainfall is sudden and heavy.

AI for Flood forecasting

Floods are one of the most common natural disasters. With a disproportionate impact in developing countries that often lack dense streamflow gauge networks, every year, floods cause thousands of fatalities worldwide, disrupt the lives of millions, and cause significant financial damages. 

According to Geoscience World, floods cause between 5,000 and 60,000 fatalities every year affecting between 95 million and 240 million people, and cause between $21 billion and $33 billion in economic damages.

The Google AI for flood forecasting combines two AI models that process diverse publicly available data sources. The first is the Hydrologic Model: it forecasts the amount of water flowing in a river. The second is the Inundation Model: it predicts what areas will be affected and how high the water level will be.

All forecasts are displayed on its platform called the Flood Hub. On it, forecasts are updated daily. And, the information is free of charge, publicly available, and can be shared over social networks. 

Flood Hub currently covers river basins in over 80 countries worldwide, providing critical flood forecasting for over 1,800 sites and, covering a population of 460M people 

Speaking for the development team, Florence Ofori told journalists that Floodhub alerts people in areas that are about to be impacted up to 7 days before disaster strikes. Google publishes the forecasts via alerts on Google Search, Maps, and Android notifications to help people access flood information.

“By warning organizations and people, we hope to empower them to act, limiting damage and loss of life”, she explained. Asked how the alerts get to the end users, She mentioned that the team is working closely with governments, the UN, and NGOs to implement and distribute flood alerts.

Dr  Jason Hickey believes that Google’s made-in-Africa AI model is ready for global adoption. “Our technology is now scalable and covers dozens of countries. In the future, we aspire to cover all areas affected by floods globally”, he added.

AI for open buildings

Africa is plagued by the consequences of poor economic planning. This is largely owing to the absence of accurate population data and geographical mappings. 

Google created a deep learning model that has been trained to determine the footprints of buildings from high-resolution satellite imagery. From there, the team created a large-scale open dataset that contains the outlines of buildings.

The project is based in Ghana, with an initial focus on the continent of Africa and new updates on South Asia, Southeast Asia, Latin America and the Caribbean. The dataset contains 1.8 billion building detections, across an inference area of 58M km².

These building footprints are useful for a range of important applications, from population estimation, urban planning and humanitarian response, to environmental and climate science.

Google researcher, Abigail Annkah said that the project is in response to complaints about traditional maps not being able to determine the actual footprints of buildings, especially in vast desert areas and locations with similar structures that are not actual buildings.

Hence, this deep learning model has been trained to determine the footprints of buildings from high-resolution satellite imagery. It has also been subjected to both omission and commission errors to ensure over 80% accuracy.

AI for Food Security

Over 70% of Africans depend on farming for their food supply and a slight disruption in environmental conditions could mean hunger for the majority. An estimated 140 million people in Africa face acute food insecurity, according to the 2022 Global Report on Food Crises 2022 Mid-Year Update. 

According to Google researcher, William Ogallo, the major reason for the lack of food security in the country is that farming is largely not mechanised and rain-led. Hence, climate change has a significant impact on farming activities and food production.

In a bid to combat food security, the Google research team in Africa has built a forecasting tool on the strength of Artificial Intelligence and climate data. The goal is to provide frequent, precise and actionable food security predictions that empower farmers to plan and redistribute farming resources intelligently.

William explained that the Google research team is passionate about helping farmers replace traditional farming methods with more effective and timely alternatives by providing accurate and frequent updates.

Concerns about responsibility: Google’s AI is for societal impact

Scholars like Harvard’s Christina Pazzanese have raised concerns about the negative potentials of artificial intelligence In the hands of profit-driven private companies:

“With virtually no government oversight, private companies use AI software to make determinations about health and medicine, employment, creditworthiness, and even criminal justice without having to answer for how they’re ensuring that programs aren’t encoded, consciously or unconsciously, with structural biases…”

Many worry whether the coming age of AI will bring new, faster, and frictionless ways to discriminate and divide at scale. In the same vein, some journalists in attendance were concerned about Google’s commitment to ethically handling the data of Africans that will be derived from the research projects. 

In response, Google developer, Perry Nelson reiterated the company’s commitment to create solutions that equally benefit all users and society. He shared with the press an overview of the principles guiding the team at Google Africa Research team as seen in the image below:

He also explained that Google is working closely with various organisations and governments to maximise efficiency and impact. These include the World Meteorological Organization and the Red Cross. These organisations are also renowned for their adherence to ethics and standards.