In 2018, Google launched its first African Artificial Intelligence, AI, research centre in Ghana in a bid to foster and
deliver innovative solutions that will benefit both Africa and the world. Since then, the centre has moved to map Africa’s built environment using satellite imagery and machine learning, built Open Buildings Datasets, advance support of Sub-Saharan African languages, reduce the harms of locusts, adapt the existing flood forecasting systems to respond to the special challenges of the continent and even tackle maternal mortality using tech. Chiemelie Ezeobi reports that a recent immersive onsite and virtual visit of African journos to the centre demonstrated AI solutions to challenges of flooding, food security, identity management and natural disasters
When Google launched its first African Artificial Intelligence (AI) research centre in Ghana in 2018, the aim according to
Yossi Matias, Vice President, Engineering & Research Google,
was to foster the continent’s growing capabilities in the field, to deliver innovative solutions that will benefit both Africa and the world, working with local universities and research centres, as well as working with policymakers on the potential uses of AI in Africa.
Google AI Research Centre
The Google Research Centre in Accra, Ghana, houses the research labs that explore how AI can be used to solve problems affecting people both locally and globally, like mapping buildings in remote locations to provide better electricity. Essentially, Google boasts that its local researchers collaborate with research teams across the globe to work on AI-based tools to create change for communities worldwide, including in various countries across Africa.
An Immersive Experience
With the comms team led by Dorothy Ooko, Communications & Public Affairs, Africa from Kenya; Communications & PR Manager, West Africa, Taiwo Kola-Ogunlade from Nigeria and
Siyavuya Madikane, Communications and Public Affairs Manager, South Africa, African journos drawn from Kenya, Nigeria and South Africa recently visited the Google Research Africa space in Accra, Ghana for an immersive firsthand experience of the ongoing projects by the Google AI research teams in collaborations with partners across the region.
In addition with home journalists from Ghana, the team explored AI products on-site with demos and case studies as was shown by Jason Hickey, who leads the Accra centre; Dr. Aisha Walcott-Bryant, Senior Staff Research Scientist;
Perry Nelson, Software Engineer, Google Ghana.
The Google team showcased the significant AI work done so far including mapping Africa’s built environment using satellite imagery and machine learning, which allowed to quadruple the number of African buildings on Google Maps, bringing the number up past 250-million.
Also, they showed the Open Buildings Dataset, which is being used by various organisations such as the UN as well as other works where the team uses machine learning and Google’s unique capabilities to advance support of Sub-Saharan African languages, reduce the harms of locusts and adapt existing flood forecasting systems to respond to the special challenges of the continent.
Why AI for Africa?
In a virtual session with Matias, he opined that since innovative solutions to problems are developed for other climes and shipped to Africa, the Google Research team in Africa is focused on developing Al solutions to solve such problems right here in Africa but with a twist- the solutions can be applied globally.
On why Africa, Yossi said the continent has become a growing tech hub of which Google contributed immensely by training thousands of tech-inclined young Africans who have gone on to found their own startups.
Also, Hickey believes that AI is a foundational and transformational technology that will provide compelling and helpful benefits to people and society through its capacity to assist, complement, empower, and inspire people in almost every field of human endeavour, especially in Africa.
In the same vein, Walcott said while they are advocating for AI, they are also practicing responsible AI to create a future where technology benefits all users and society while working alongside diverse collaborators, partners and external contributors to make AI more transparent, fair, and useful to diverse communities, as well as constantly improve the reliability and safety of the entire AI ecosystem.
Corroborating, Google’s Perry Nelson, while addressing concerns about responsibility, he said the company is working closely with various governments and even organisations in this regard. To maximise efficiency and impact without endangering the society, he said Google works with World Meteorological Organisation and the Red Cross, renowned for being stickler to standards and ethics.
AI for Weather Forecasting
At the centre, the project lead for AI for Weather Forecasting, Emmanuel Brempong talked about AI for weather forecasting and how Google’s weather forecasting tool has been built on machine learning models that can precisely estimate current precipitation conditions. Already, it has been used to provide early warnings in Africa, especially before any sudden and heavy rainfall.
Corroborating afterwards, Google’s Walcott said the potential such accurate weather forecasts have on agriculture that rely on weather for decision making, is enormous.
AI Dataset for Open Buildings
With an initial focus on Africa and new updates on South Asia, Southeast Asia, Latin America, the Caribbean and 12 other countries, this dataset contains 1.8 billion building detections, across an inference area of 58M km2.
To do this, Google created a deep learning model trained to determine the footprints of buildings from high-resolution satellite imagery. With this, they went on to create a large scale open dataset that contains the buildings outline or footprint, which are useful for population estimation, urban planning and humanitarian response, to environmental and climate science.
According to Google researcher, Abigail Annkah, the team did this in response to issue of traditional maps not being able to determine the actual footprints of buildings, especially in vast areas and locations with structures that are not even buildings. This is because even with satellite imagery, it can be difficult to map buildings in remote locations and when these buildings go unmapped, it can make things like planning infrastructure difficult.
AI For Flood Forecasting
With the weather forecast working, the Google team spread its tentacle specifically to Flood Forecasting. This is because in Africa, floods are one of the most common natural disasters and even though it’s expected, it never fails to leave massive destruction in its wake as the respective governments always seem ill prepared to tackle it.
In this regard, the Google AI for Flood Forecasting combines the Hydrologic and Inundation Models. While the first forecasts the amount of water flowing in a river, the second predicts what areas will be affected and how high the water level will be.
To make this forecasting more effective, the Flood Hub was borne as a platform for such forecasts to be updated daily, free of charge and publicly available.
Currently, this covers river basins in over 80 countries worldwide where it provides critical flood forecasting for over 1,800 sites.
According to Google’s Florence Ofori, this Flood Hub is so critical as it alerts people in potential flood areas up to seven days before it even happens. And fo ensure its wide usage, these alerts are published on Google Maps, Google Search, Maps, and even Android notifications.
Google also partnered governments of the days, the United Nations, and non governmental organisations to distribute these flood alerts.
Essentially, this Google’s lifesaving technology- Flood Forecasting Initiative is helping residents stay safe and give governments time to prepare. These AI models predict when and where riverine floods will occur in 80 countries worldwide, including 23 in Africa.
Improving Maternal Health Outcomes with Ultrasound
Another score for the Google team is the Ultrasounds, which can be crucial for identifying potential complications during pregnancy. In recent years, sensor technology has evolved to make ultrasound devices significantly more portable and affordable.
Globally, Google said it has been been working on building AI models that can read ultrasound images and provide important information to healthcare workers. “In Kenya, for instance, we are partnering with Jacaranda Health to help improve our ultrasound AI technology, with a focus on using handheld ultrasound devices that don’t need to be attached to larger machines.
“This can help people who aren’t trained to operate traditional ultrasound machines to acquire and interpret ultrasound images and triage high-risk patients, simply by sweeping the handheld probe across the mother’s belly”.
Project Relate
Project Relate helps people with non-standard speech make their voices heard. The Android app uses AI research to help people with non-standard speech communicate more easily and after recording 500 phrases, users receive a personalised speech recognition model.
This app can transcribe speech into text; use a synthesised voice to repeat what the speaker has said; and engage Google Assistant to complete tasks, such as asking for directions, playing a song or turning on the lights.
Read Along
Another way the Google team have utilised AI is with Read Along through which they are teaching reading to children worldwide. Created based on the impact of COVID-19 on reading, the Read Along Google’s AI-based reading tutor app and website, is helping to increase child literacy.
On the other hand, Diya, the in-app reading buddy, listens to the speaker reading aloud, offering support when they struggle, and rewarding them when they do well.
According to Google, over the past three years, more than 30 million kids have read more than 120 million stories on Read Along.
AI for Food Security
Locust infestations can have a devastating effect on food crops. So through collaborations with AI-product focused company InstaDeep and the Food and Agriculture Organisation (FAO) of the United Nations, the Google team at the Google AI Center in Ghana is helping to better detect locust outbreaks and enable farmers to implement control measures.
The AI Center team is also working on building a model that forecasts locust breeding grounds using historical data from the FAO and environmental variables like rainfall and temperature.
These are some of the ways Google is using AI to ensure food security especially as the challenge of food security in Africa is a burgeoning one
Noting that lack of food security in Africa is because of unmechanised faming and Google researcher, William Ogallo, said it was in a bid to combat food security that the Google research team in Africa built a forecasting tool on the strength of Artificial Intelligence and climate data provide frequent, precise and actionable food security predictions that empower farmers to plan and redistribute farming resources intelligently.
With this tool, traditional farming methods will be replaced by enhancing the frequency and accuracy of forecasts, to mitigate the risks of food crises before they escalate using the benefits of AI-driven agility and efficiency.
Summarily, at the end of the immersive experience, it was obvious how the Google Research team in Africa is not just focused on developing Al solutions that will solve African problems through data, but also those that will have global reach, as are already happening with the flood forecasting and weather forecasting.