In January 2018, CIFAR and the Brookfield Institute for Innovation + Entrepreneurship (BII+E) formed a partnership to develop and host five AI Futures Policy Labs aimed at generating greater awareness of the long-term implications of AI. These workshops were also designed to build capacity among emerging policy leaders to produce agile AI policy in Canada. Between June and September 2018, CIFAR and BII+E hosted two labs in Toronto, Ontario and Edmonton, Alberta, engaging a total of 41 emerging policy leaders from the public, private, academic, and not-for-profit sectors.
On October 18, 2018, BII+E and CIFAR hosted the third AI Futures Policy Lab in Vancouver, British Columbia. This event brought together 22 emerging policy leaders from Vancouver and Victoria with the aim of:
- Increasing the capacity of future public service leaders to understand the policy implications of AI;
- Cutting through the myths and hype surrounding AI to provide policymakers with a direct line of sight into the AI sector, current capabilities, and potential applications;
- Facilitating early thinking around appropriate government responses to emerging AI technologies.
The design of this lab was largely informed by participant feedback gathered at the Edmonton lab in September 2018. Alongside the AI 101 talk and guest speaker, an AI Policy 101 talk was incorporated into the morning’s agenda. This provided participants with a better understanding of the current policies and initiatives in place to address AI’s societal implications, both in Canada and abroad. In the afternoon, each group was provided with an example of a current AI application that would serve as the basis of their analysis and discussions. Potential future scenarios that had been used in the previous labs were discarded due to previous feedback from former participants, which indicated lack of flexibility in future-thinking discussions. This change was made to enable participants to imagine and discuss their own scenarios with greater latitude. During the final session of the day, participants were supplied with a template that prompted them to collectively produce policy recommendations related to the scenario they explored. The agenda developed for the day is provided in appendix A.
Case Study Policy Domains
Prior to the lab, four profiles of current AI applications were developed, each associated with a specific domain. This included housing, legal, education, and health. Groups of 4-5 participants, led by a facilitator, were each assigned a different domain to focus on during the afternoon sessions. These are listed below.
AI is impacting the housing sector in multiple ways, from smart-home devices like Nest to intelligent tools that help to curb energy use, and services that even act as the middle-man between landlords and tenants. Advancements in this domain afford residents with potential benefits, but also create challenges regarding privacy and safety in a domestic environment. Within this domain, participants examined Naborly (appendix B), a tenant screening application that generates risk scores to help landlords make smarter letting decisions.
The legal sector is being impacted by recent developments in AI and machine learning capabilities that have enabled applications to automate legal research, due diligence processes, contract review and management practices, and help to predict legal outcomes. Participants within this domain were given the chance to explore the policy impacts of ROSS intelligence (appendix C), an artificially intelligent legal research tool that applies natural language processing to increase lawyer’s ability to identify relevant information.
There is vast potential for AI to transform education in ways that make learning more accessible, provide personalized curriculum, and support educators in delivering content. Participants in this group analyzed Nestor (appendix D), an artificially intelligent class assistant that uses machine learning algorithms and advanced facial recognition to analyze the attention of students listening to online lectures.
Advancements in AI capabilities hold enormous opportunities for delivering more efficient health care services in areas such as diagnosis, health monitoring, and treatments. However, this also raises challenges related to patient privacy and discrimination. Participants within this group explored InnerEye (appendix E), a research initiative led by Microsoft that applies computer vision and machine learning algorithms to automatically analyze three-dimensional medical CT (computer tomography) and MR (magnetic resonance) images to identify tumours.
Read the full report here
In partnership with the Brookfield Institute for Innovation + Entrepreneurship
(BII+E), this project has been designed to help emerging policy leaders across Canada understand and respond to the opportunities and challenges that accompany the rapid development of artificial intelligence.