Presenter: Rashid Mushkani
Affiliation: Doctoral Candidate, University of Montreal
Conference: Innovate for Cities 2024
Public spaces in cities were primarily designed in the mid-20th century, often not accommodating the diverse and evolving needs of today’s populations. Challenges such as global warming and increased urban diversity necessitate a reevaluation of how inclusive these spaces are for all community members, including the elderly, disabled, minority groups, women, and LGBTQ+ individuals.
A predictive AI model was developed to assess the qualities of public spaces. This model was trained on approximately 60 data points and achieved around 90% accuracy in predicting inclusivity based on various spatial elements such as sidewalk construction and surrounding buildings.
A generative AI model was created to produce direct conceptual designs for public spaces. This model generates designs that are fine-tuned to the context of Montreal, facilitating the visualization of inclusive and functional urban environments.
Workshops were conducted with 20 participants from diverse backgrounds, including elderly, disabled, minority, women, and LGBTQ+ communities. These sessions helped formulate AI prompts and ensure that the models reflect the perspectives of different user groups.
The predictive AI model successfully identified inclusive and exclusive spaces, providing heat maps that highlight areas in Montreal needing improvement. The generative AI model produced nuanced visualizations of public spaces that cater to diverse community needs. These tools offer valuable insights for urban planners to make informed decisions, enhancing the inclusivity and functionality of public spaces.