Policy Recommendations

The AI4HealthyCities dialogues between São Paulo and Quebec were consolidated into policy recommendations on how to successfully implement sustainable data, digital and AI led innovations for population health.

Who are the summary findings and recommendations from AI4HealthyCities for?

The recommendations are designed for anyone directly or indirectly involved with or enabling decision-making related to urban health and wellbeing, such as City authorities from different sectors (health, innovation, tech, communication, finance etc.), citizens, civil society, academia and private sector players.

Create a multistakeholder forum

Partnerships and Multistakeholder Collaborations

  • Create a permanent multistakeholder forum to strengthen the relationship between the Government of Québec and the State of São Paulo.
  • Implement multiple collaboration formats, such as official long-term partnership agreements, bilateral agreements, a steering committee to oversee the deployment of AI technologies, knowledge sharing platforms, and mutual milestones.
  • Develop a clear roadmap and methodology on how to create and maintain a functional multistakeholder network. methodology on how to create and maintain a functional multistakeholder network.

Engage with Civil Society and Non-profit Organizations

Partnerships and Multistakeholder Collaborations

  • Foster learning opportunities and build a common agenda with various stakeholders to increase diversity, equity, and inclusion in the development and implementation of AI technologies to eliminate health disparities.
  • Establish a regular communication channel with civil society and non-profit organizations through workshops, public hearings, public consultations, studies, grants, and online engagement.
  • Include underrepresented communities as a means of avoiding potential harm caused by irresponsible AI (such as Indigenous peoples, Black people, people of colour, and LGBTQIA+ people).

Foster Local Innovation

Funding and Policy

  • Create incentives and opportunities for R&D groups to partner with the private sector and civil society organizations.
  • Fund challenges to support local companies providing responsible and ethical AI solutions in health, and that follow diversity, equity and inclusion policies.
  • Provide seed funding for actors of the local AI ecosystem committed to developing responsible and ethical AI and help integration into the larger national AI ecosystem.
  • Avoid the “pilot trap” — when governments are eager to develop pilot projects but fail to commit resources to scale them — by establishing mid-term views and mechanisms for support beyond the initial stage.

Improve Data Literacy

Capacity Building

  • Partner with specialists, public libraries, cultural centers, schools and universities to offer data literacy courses free of charge and open to the public, both in-person and online.
  • Include data and media literacy in the curriculum of public and private schools.
  • Design and promote campaigns to enhance awareness of the importance of data and media literacy.

Invest and Develop Responsible, Ethical, Accountable AI Solutions

Ethics and Rights

  • Invest in education, training and debates addressing the ethics, responsibility, accountability, policy, and governance of AI.
  • Integrate human and social sciences training in the curricula of data scientists and other technology developers.
  • Hire people committed to diversity, equity, and inclusion to develop, deploy and evaluate AI solutions.
  • Include civil society and non-profit organizations in the debate around ethical and responsible AI to learn together.
  • Fund interdisciplinary, multistakeholder teams to develop open, accessible guidelines to assess responsibility and accountability of AI solutions and encourage the public and private sectors to adopt these guidelines when buying selling products and services.
  • Adopt data governance and protection frameworks that promote public interest and digital rights.
  • Implement AI and data use registries to enhance transparency.
  • Include underrepresented communities (e.g. indigenous peoples, Black people, people of color, and members of the LGBTQIA+ community).
  • Ensure funding guidelines prioritize projects committed to diversity, equity and inclusion.

Sustain Public, Open Datasets

Infrastructure and Datasets

  • Build ethical, public, and open datasets that comply with regulatory frameworks of data governance and protection, digital rights, and guidelines from the openness community.
  • Consult with data scientists and other specialists on priorities and expectations for open/public datasets to improve data quality and interoperability.
  • Adopt robust data governance and protection frameworks.
  • Create incentives for contributions to open datasets, through collaborations and partnerships between governments, industry, academia and civil society.
  • Invest in data security, protection, and privacy.

Develop a Platform to Share Resources and Knowledge

Partnerships and Multistakeholder Collaborations

  • Develop an AI and health platform for knowledge sharing between regions and exchange of best practices, lessons learned, failures not to be repeated, works in progress, and international perspectives.
  • Shared resources and materials should prioritize openness (open licensing, access, data, codes, educational resources) to ensure equity and access to knowledge for all whenever possible.

Invest in and Promote Research and Development (R&D)

Funding and Policy

  • Increase long-term sustainable R&D funding across sectors (academia, industry and civil society) independent of current political leadership.
  • Promote and fund regular, public conferences, symposiums and other opportunities to showcase research findings, research-action projects, and evidence-based policy-making.
  • Foster interdisciplinary support for grant applications (for example, a combination of STEM, Humanities, and Social Sciences) and set guidelines for diversity, equity and inclusion that projects should follow to get funded.
  • Fund mentorship programs to facilitate regular exchange on evidence-driven and science-informed initiatives between experts, advocates, and policymakers, encouraging openness about successes and limitations.
  • Have independent publicly funded labs dedicated to evaluating the ethical aspects of AI research projects.

Train and Retain the Workforce

Capacity Building

  • Develop interdisciplinary, AI-focused capacity building programs and workshops for public servants that address potentialities, challenges and issues related to the design and implementation of AI initiatives, especially when related to health and healthcare.
  • Gather support from C-executives and managers, ensuring workers have sufficient time and support for training.
  • Foster AI knowledge in nontechnical fields (policy, legal, sociology, urban planning). This may include the development of interdisciplinary AI-focused capacity building programs and workshops for public servants. Endorsement from executives and managers should be ensured.
  • Create programs and funding opportunities to recruit and retain leading global talent in the field of AI, and AI and health, and increase the number of research-focused employees.
  • Design strategies to overcome language barriers and ensure diversity, equity and inclusion.
  • Support organizations through guidance and tools to engage a diverse group of citizens and patients in digital health projects.

Design Responsible AI Assessment and Acquisition Tools

Regulatory Framework

  • Collaborate with multistakeholder specialists to set standards, methodologies, and guidelines for AI acquisition and deployment that include bias, risk, or harm assessment mechanisms.
  • Develop tools to evaluate sustainability of AI solutions in the short, mid, and long- term.
  • Design clear norms and guidelines for AI procurement and Public-Private Partnerships in a multistakeholder dialogue.
  • Set up an independent multistakeholder entity (steering committee), to oversee projects and actors across the ecosystem, and to set standards and metrics for monitoring and evaluating impact.

Fund Connectivity and Broadband Access Programs

Infrastructure and Datasets

  • Fund programs to improve internet universalization and broadband access in public spaces and in households.
  • Ensure these programs follow best practices and regulatory frameworks for data governance and protection.
  • Design these programs in collaboration with stakeholders outside of government and industry to guarantee public interest connectivity and compliance with digital
    rights frameworks.

Prioritize Responsible and Ethical Telemedicine

Telemedicine

  • Develop and integrate responsible telemedicine options to improve access to
    healthcare for patients and professionals that most need it.
  • Design reliable follow-up processes to ensure predictive or preventative
    telemedicine solutions are linked to an effective clinical response when needed.
  • Build and publish protocols and guidelines for the responsible and ethical
    deployment of telemedicine in consultation with specialists from different fields
    (medicine, law, public policy, data science) and practitioners (health, IT, social
    services, legal services).
  • Follow regulatory frameworks for data governance and protection to ensure patient data collection and usage in digital devices is compliant with personal rights and avoids widening health disparities.
  • Invest in interoperability and integration.
  • Establish models and methodologies for monitoring and impact evaluation to ensure innovations demonstrate their value in the short, mid and long term.
  • Consult with end-users to make technology user-friendly, needs-based, and
    accessible.

Create a multistakeholder forum

Partnerships and Multistakeholder Collaborations

  • Create a permanent multistakeholder forum to strengthen the relationship between the Government of Québec and the State of São Paulo.
  • Implement multiple collaboration formats, such as official long-term partnership agreements, bilateral agreements, a steering committee to oversee the deployment of AI technologies, knowledge sharing platforms, and mutual milestones.
  • Develop a clear roadmap and methodology on how to create and maintain a functional multistakeholder network. methodology on how to create and maintain a functional multistakeholder network.

Develop a Platform to Share Resources and Knowledge

Partnerships and Multistakeholder Collaborations

  • Develop an AI and health platform for knowledge sharing between regions and exchange of best practices, lessons learned, failures not to be repeated, works in progress, and international perspectives.
  • Shared resources and materials should prioritize openness (open licensing, access, data, codes, educational resources) to ensure equity and access to knowledge for all whenever possible.

Engage with Civil Society and Non-profit Organizations

Partnerships and Multistakeholder Collaborations

  • Foster learning opportunities and build a common agenda with various stakeholders to increase diversity, equity, and inclusion in the development and implementation of AI technologies to eliminate health disparities.
  • Establish a regular communication channel with civil society and non-profit organizations through workshops, public hearings, public consultations, studies, grants, and online engagement.
  • Include underrepresented communities as a means of avoiding potential harm caused by irresponsible AI (such as Indigenous peoples, Black people, people of colour, and LGBTQIA+ people).

Invest in and Promote Research and Development (R&D)

Funding and Policy

  • Increase long-term sustainable R&D funding across sectors (academia, industry and civil society) independent of current political leadership.
  • Promote and fund regular, public conferences, symposiums and other opportunities to showcase research findings, research-action projects, and evidence-based policy-making.
  • Foster interdisciplinary support for grant applications (for example, a combination of STEM, Humanities, and Social Sciences) and set guidelines for diversity, equity and inclusion that projects should follow to get funded.
  • Fund mentorship programs to facilitate regular exchange on evidence-driven and science-informed initiatives between experts, advocates, and policymakers, encouraging openness about successes and limitations.
  • Have independent publicly funded labs dedicated to evaluating the ethical aspects of AI research projects.

Foster Local Innovation

Funding and Policy

  • Create incentives and opportunities for R&D groups to partner with the private sector and civil society organizations.
  • Fund challenges to support local companies providing responsible and ethical AI solutions in health, and that follow diversity, equity and inclusion policies.
  • Provide seed funding for actors of the local AI ecosystem committed to developing responsible and ethical AI and help integration into the larger national AI ecosystem.
  • Avoid the “pilot trap” — when governments are eager to develop pilot projects but fail to commit resources to scale them — by establishing mid-term views and mechanisms for support beyond the initial stage.

Train and Retain the Workforce

Capacity Building

  • Develop interdisciplinary, AI-focused capacity building programs and workshops for public servants that address potentialities, challenges and issues related to the design and implementation of AI initiatives, especially when related to health and healthcare.
  • Gather support from C-executives and managers, ensuring workers have sufficient time and support for training.
  • Foster AI knowledge in nontechnical fields (policy, legal, sociology, urban planning). This may include the development of interdisciplinary AI-focused capacity building programs and workshops for public servants. Endorsement from executives and managers should be ensured.
  • Create programs and funding opportunities to recruit and retain leading global talent in the field of AI, and AI and health, and increase the number of research-focused employees.
  • Design strategies to overcome language barriers and ensure diversity, equity and inclusion.
  • Support organizations through guidance and tools to engage a diverse group of citizens and patients in digital health projects.

Improve Data Literacy

Capacity Building

  • Partner with specialists, public libraries, cultural centers, schools and universities to offer data literacy courses free of charge and open to the public, both in-person and online.
  • Include data and media literacy in the curriculum of public and private schools.
  • Design and promote campaigns to enhance awareness of the importance of data and media literacy.

Design Responsible AI Assessment and Acquisition Tools

Regulatory Framework

  • Collaborate with multistakeholder specialists to set standards, methodologies, and guidelines for AI acquisition and deployment that include bias, risk, or harm assessment mechanisms.
  • Develop tools to evaluate sustainability of AI solutions in the short, mid, and long- term.
  • Design clear norms and guidelines for AI procurement and Public-Private Partnerships in a multistakeholder dialogue.
  • Set up an independent multistakeholder entity (steering committee), to oversee projects and actors across the ecosystem, and to set standards and metrics for monitoring and evaluating impact.

Invest and Develop Responsible, Ethical, Accountable AI Solutions

Ethics and Rights

  • Invest in education, training and debates addressing the ethics, responsibility, accountability, policy, and governance of AI.
  • Integrate human and social sciences training in the curricula of data scientists and other technology developers.
  • Hire people committed to diversity, equity, and inclusion to develop, deploy and evaluate AI solutions.
  • Include civil society and non-profit organizations in the debate around ethical and responsible AI to learn together.
  • Fund interdisciplinary, multistakeholder teams to develop open, accessible guidelines to assess responsibility and accountability of AI solutions and encourage the public and private sectors to adopt these guidelines when buying selling products and services.
  • Adopt data governance and protection frameworks that promote public interest and digital rights.
  • Implement AI and data use registries to enhance transparency.
  • Include underrepresented communities (e.g. indigenous peoples, Black people, people of color, and members of the LGBTQIA+ community).
  • Ensure funding guidelines prioritize projects committed to diversity, equity and inclusion.

Fund Connectivity and Broadband Access Programs

Infrastructure and Datasets

  • Fund programs to improve internet universalization and broadband access in public spaces and in households.
  • Ensure these programs follow best practices and regulatory frameworks for data governance and protection.
  • Design these programs in collaboration with stakeholders outside of government and industry to guarantee public interest connectivity and compliance with digital
    rights frameworks.

Sustain Public, Open Datasets

Infrastructure and Datasets

  • Build ethical, public, and open datasets that comply with regulatory frameworks of data governance and protection, digital rights, and guidelines from the openness community.
  • Consult with data scientists and other specialists on priorities and expectations for open/public datasets to improve data quality and interoperability.
  • Adopt robust data governance and protection frameworks.
  • Create incentives for contributions to open datasets, through collaborations and partnerships between governments, industry, academia and civil society.
  • Invest in data security, protection, and privacy.

Prioritize Responsible and Ethical Telemedicine

Telemedicine

  • Develop and integrate responsible telemedicine options to improve access to
    healthcare for patients and professionals that most need it.
  • Design reliable follow-up processes to ensure predictive or preventative
    telemedicine solutions are linked to an effective clinical response when needed.
  • Build and publish protocols and guidelines for the responsible and ethical
    deployment of telemedicine in consultation with specialists from different fields
    (medicine, law, public policy, data science) and practitioners (health, IT, social
    services, legal services).
  • Follow regulatory frameworks for data governance and protection to ensure patient data collection and usage in digital devices is compliant with personal rights and avoids widening health disparities.
  • Invest in interoperability and integration.
  • Establish models and methodologies for monitoring and impact evaluation to ensure innovations demonstrate their value in the short, mid and long term.
  • Consult with end-users to make technology user-friendly, needs-based, and
    accessible.

Methodology

The twelve policy recommendations build upon key themes. Each policy recommendation features a list of actions to accomplish their goals and relevant enablers that should be involved in the process. There is no specific order, and more than one policy recommendation can happen simultaneously.

The following policy recommendations will help stakeholders around the world build and sustain the ecosystem of companies, researchers, and leading thinkers on the potential of using AI to improve urban population health, while respecting regulatory frameworks and ensuring a responsible, ethical approach to the deployment of AI in urban spaces. 

They were developed from close analysis of the successes, challenges, priorities, and needs shared by multisectoral experts from both Québec and São Paulo in the use of AI for health. They reflect what has worked well - and what hasn’t - to help policymakers and anyone working to use AI to improve urban population health.

Other cities around the world can apply the following policy recommendations in their own contexts. But engaging with other regions closely can lead to the development of more policies that respond to other challenges and opportunities in the field.

Download our 50 pages report

The AI4HealthyCities: Recommendations for Using AI to Transform Urban Health report outlines the key findings of the 2021 AI4HealthyCities dialogues, highlighting case studies on the use of AI for improving urban health in addition to the 12 policy recommendations. The report also summarizes AI4HealthyCities’ scalable and replicable engagement model as a call to action for governments around the world to join the conversation.

Download Report