A high quality of life for all: just urban planning requires data diversity and inclusive AI

Visitors to Nigeria’s capital Lagos are unlikely to avoid the city’s bustling and noisy traffic. Africa’s second largest city is no different to other major cities in that its infrastructure is failing to keep pace with growth. The impacts of climate change are taking their toll. Energy and water requirements are rising. At the same time, citizens want an efficient administration – and to feel safe in public spaces.

Smaller urban centres are growing too, not just metropolises like Lagos. By 2050, two thirds of the global population will live in cities. Here, the problems are felt in a context characterised by a high population density. Poverty, health risks, waste and supply problems coexist side by side. Cities already account for more than 70 per cent of global CO2 emissions. Whether the 2030 Agenda and the Paris Agreement are implemented or not will be decided in cities, which is why German development cooperation is right to promote sustainable urban development. Doing so is an efficient step since even minor changes in cities affect lots of people simultaneously and help to mitigate climate change and combat poverty.

In this context, urban planners are relying increasingly on data and artificial intelligence, whether it be to control traffic, improve waste collection or make public administration more accessible. The Deutsche Gesellschaft für Internationale Zusammenarbeit (GIZ) GmbH is also working on solutions to make cities in Africa, Asia and Latin America smarter. With many cities facing similar challenges, it is good that mayors are not alone. They share their views with one another at a global level, for example within the C40 network of international cities. This allows German and European cities to benefit from the experiences of Lagos, Nairobi or Cape Town as well – and vice versa.

Data diversity for equitable smart cities

One thing that is true for any city is that ‘smart’ is only effective if technological advancements are inclusive of all who live there. One reason why this does not occur is because population groups that are already structurally disadvantaged often fail to show up in datasets. This applies as much to older people and those with disabilities as it does to women and girls. Working people from poor city districts are also disadvantaged as they are the ones who, ultimately, often have to tolerate long journeys to work. If they are not included in the data used to plan cities, they will not be taken into consideration when decisions are made.

Lagos serves as an inspiration in this respect. Here – and in three other African cities – researchers surveyed more than 700 women and men on the mode of transport they use and why. They also observed mobility patterns. In Lagos, this was the first large-scale comparative data survey on gender and mobility – and it was commissioned by GIZ. The result shows that women and girls use public spaces differently to men. Women travel by bus more frequently, but at different times. They change more often, for instance because they first drop their children off, before doing some shopping and then visiting family. They walk more but also, unfortunately, are more likely to suffer sexual harassment. These data are key for the Lagos Metropolitan Area Transport Authority, which has begun implementing projects that make transport safer, in particular for women and girls.

Highlighting the needs of marginalised groups

Algorithms can also discriminate – if they are based on biased or incomplete data. If urban planners rely on AI to recommend where to build the next rescue station or school, they might make the wrong decision, for example because AI fails to identify underserved areas as a priority due to incomplete data.

The example of Mexico City is one to follow. There, the city government, a non-governmental organisation (NGO) and Mexico’s Women’s Ministry surveyed around 10,000 women on why they do not work and what needs to change. The data show that nurseries and care facilities are often much too far away. Anybody living on the outskirts of Mexico City travels several hours on public transport just to take their children to a nursery and get to work themselves. Little time remains for paid work – which translates into little money at the end of the month, if the effort is even worth it.

With support from GIZ, the data were made publicly accessible on a platform, with a map clearly showing where the gaps are. The Women’s Ministry uses the platform to plan the construction of new care facilities in Mexico City. Its aim is for everyone to be within 15 minutes’ walk of such a facility. This places care work on a broader basis and relieves the burden on women. An algorithm on the platform automatically recommends suitable locations.

Diverse teams for inclusive AI

Examples such as Mexico City and Lagos offer encouragement. The more diverse the people who shape these cities are, the more inclusive the cities, in particular, become. That is why more marginalised groups, including women, are required in the data and AI sectors. GIZ is working worldwide with its partner organisations to train these groups. In South Africa, Ghana and Rwanda, for instance, there is the AI and Data Science Bootcamp, an AI and data training course that pays particular attention to the everyday realities of marginalised groups.

The cities of tomorrow are intelligent and progressive. And they allow everyone to take part in public life in equal measure. For this to succeed, urban planning requires diverse teams. Marginalised groups need to feature in the data sets and their downstream AI applications. And international city networks are needed so that good smart city initiatives can be replicated worldwide.

Published on 29 May 2024: Just urban planning needs data diversity and inclusive AI – Tagesspiegel Background (in German)