Smart tech to tackle homelessness in London, ON
In a joint Data Governance lab with Evergreen and MaRS, cities are exploring what role data governance may have in using smart technologies for the betterment of city residents.
Published on March 11, 2020
Evergreen and MaRS are prototyping ways to use data and tech for betterment of the city and its residents
The era of the smart city is now.
Powerful technological advances are accelerating the pace of change through sophisticated methods of data analysis that, in turn, are steadily changing how we work, learn, play, live, and connect with others. While IT and AI developments can hold the keys to transform how Canadian cities render services for the benefit of all, cities are struggling with how to effectively harness these technological capabilities while balancing the concerns to public trust and privacy, as well as the ethical issues arising from reducing the lived experiences of residents into sets of 1s and 0s.
A central concern is data governance, the process of managing the integrity, use, security and availability of data in the context of our ever expanding uses of data toward maximizing efficient allocation of public funds vis-à-vis quality of life improvements at the municipal level. This is a key consideration as the City of London, ON develops data governance frameworks to support implementation and scaling of the City’s Homeless Individuals and Family Services Information AI model, built by the Homeless Prevention Division to predict the probability of chronic homelessness of individuals in shelter systems and aid in resource prioritization.
Homelessness is an urgent, expensive issue in Canada, and access to adequate housing is a human right. By more recent estimates, up to 235,000 people in Canada experience homelessness in any given year. The federal government embarked on the first-ever National Housing Strategy as a way to address the increasingly compelling issue of chronic homelessness, with upwards of $2.2 billion investment into creating affordable housing. Municipalities are in critical need of solutions to ease the costs associated with homelessness on their budgets and residents.
The new AI model is an exciting tool to assist in solving homelessness but it comes with significant caveats. The algorithm collates data from multi-sector city services to predict when someone is at high risk of chronic homelessness and consequently triggers triaged intervention resulting in significant cost prevention and quality of life improvements. However, a key barrier to its adoption is the ‘black box’ nature of AI. ‘Black-box’ refers to an AI model’s opaqueness with respect to understanding ‘why’ the AI model makes the predictions it does. While it is very good at predicting chronic homelessness, it is important to know why it makes the predictions it does to satisfy due diligence and accountability requirements with respect to public sector resource deployment.
So, how do we develop meaningful data governance to address these considerations?
Evergreen, in partnership with the MaRS Discovery District, has teamed up with London as one of three cities who are ideating, prototyping, and testing ways to support the use of powerful data and information technologies for the betterment of the city and its residents. The Data Governance Lab, a Future Cities Canada program, is also working with Mississauga and Calgary on other projects that will bring their data governance initiatives to life with deeper and more nuanced understandings of the structures required to operationalize them in real world environments.
In London, this meant engaging key members of city service units to gain understanding of their needs in explaining what a black box AI model would entail. This process is allowing London to begin forming the basic considerations that would enable this project to achieve minimum viability as well as outlining what a proactive policy on automated decision making systems might look like.
The Data Governance Lab is in its final stages of wrapping up the first iteration with the learnings being compiled and shared far and wide in the next couple months. If you’re interested in learning more about this issue, and the work being done by the other labs, sign up to the Future Cities Canada newsletter!