top of page

How Can AI Help Tackle Mobility Poverty?

Updated: Aug 24, 2021


Image: City of Amsterdam



In the past decades, Artificial Intelligence has been the leading force of technological evolution. From the convenience of personalised movie recommendations to the life-saving improvements of automated medical diagnoses, our world is constantly being disrupted, and AI is the main reason behind it. But algorithms do not stop at providing better solutions to existing problems. They have the potential to provide us with a deeper understanding of our society, discover previously unnoticed inequalities, and provide the tools to accurately measure and eventually tackle them.

One of these problems is mobility poverty.


What is mobility poverty?

In today's fast-paced world, the ability to go from one place to the other quickly, cheaply, and without effort is often taken for granted. However, not everyone can afford or is able to enjoy this benefit. On top of that, the shift of city planning towards utility rather than inclusivity, combined with the increasing rental prices, has led to mobility becoming less of a right and more of a privilege.


Not having access to sufficient mobility leads to certain groups of people being limited from participating in daily activities, such as labour, education, and social events. This phenomenon is known as mobility poverty. It mainly affects low-income people, unemployed job seekers, the elderly, and citizens without a driver's license.


While The Netherlands is the country with the highest bicycle usage in the world, it is not completely immune to mobility poverty, as still only 1/4 of the daily trips are being done by bicycle.


What are the challenges regarding mobility poverty?

The main difficulty in tackling mobility poverty is that there is no standard way of measuring it. While it is evident that it is most prevalent in disadvantaged groups, there is limited insight into the size and composition of these groups. Unlike income poverty, where a monetary value can be used as the cutoff point, mobility is more complicated. A person's moving patterns are influenced by their accessibility to different amenities, but also by personal preferences on residing location and mode of transport. For example, some people might prefer living near their workplace and others near the most social activities. Thus, being less mobile does not necessarily mean that one suffers from mobility poverty.


Another challenge that both researchers and policymakers face is recommending interventions to eliminate mobility poverty. Even if the number of people suffering from it was known, the actions needed to tackle it vary significantly by country and city. Furthermore, measuring the impact of policy is difficult and mostly relies on educated guesses. Some researchers, for example, argue for more public transport subsidies towards the disadvantaged groups and for increased frequency of buses to and from rural communities, while others argue for a complete overhaul of transport planning towards a more just and equitable system.


How can AI help overcome these challenges?

Artificial Intelligence can be a catalyst in understanding and overcoming the challenges outlined above. The vast amount of movement data collected nowadays by smartphones, public transport cards, and vehicle GPS can be employed to analyse how people move within the city.


To address the measurement challenge, mobility data can be combined with socio-economic data and then be used as an input to network clustering algorithms. These algorithms can extract mobility patterns and identify clusters of people who move similarly across different socio-economic properties. This analysis can potentially lead to detecting outlier minorities, whose movement significantly differs from the majority.


Tackling the policy intervention challenge would require a shift of city planning towards accessibility, rather than utility. Nowadays, businesses tend to use mobility flow prediction models to determine the spots in a city that will attract the most possible customers. These models can be adjusted to optimise for fairer accessibility instead and be used by city planners for deciding where to build parks, theatres, community centres, etc.


Finally, Amsterdam can serve as a case study on how extensive and connected bicycle networks can be a solution against the rise of mobility inequalities in big cities. Through the use of simulations, other cities can analyse if a similar transportation planning practice can lead to more accessible and fairer mobility.


Conclusion

One of the main goals of the United Nations Sustainable Development Goals is to provide access to affordable, accessible, and sustainable urban systems for all, especially those who need it the most. Artificial intelligence can be an important ally in our efforts to reach that goal by the end of this decade. In the Civic AI lab of the University of Amsterdam, we strive to make this possible.

Comments


bottom of page