Furthering AI Research & Ethics – Towards a Digital Rights Agenda for Ontario
As technology brings about disruption across sectors, the provincial and federal governments kick off a series of initiatives to address its increasingly important role in society. Summer 2018 was vibrant on the technology front. In June, Canada’s Prime Minister hosted the G7 Summit where world leaders committed to develop a “common vision of human-centric AI”. Shortly thereafter, Canada’s Minister of Innovation, Science, and Economic Development announced the launch of the National Digital and Data Consultations with a view to discussing how the country should drive technology innovation and handle the legal and ethical issues in that regard.
With Canada well-positioned as an innovation leader, governments, policymakers, think tanks, profit and non-profit organizations alike may want to get acquainted with what is happening at each stage of the development of our digitalized society.
In Ontario, the Law Commission (the “LCO”) and the Mozilla Foundation (“Mozilla”) organized a roundtable to develop a digital rights agenda for the province and beyond. This 4-phase project involved: (1) research; (2) participant interviews; (3) identifying themes; and (4) facilitating discussion. The phases in concert aimed at identifying commonly shared concerns and potential legal issues to ensure the creation of ethics-driven guidelines for AI. At this stage, the goal was to lay out the foundations for more concrete proposals and discussions for later consultation processes.
Here’s a snapshot of the six general law reform themes or priorities outlined in the “Digital Rights, Digital Society Roundtable Report” (the “Report”) by the LCO and Mozilla:
1) Algorithmic Accountability
Algorithms increasingly drive efficiency and speed in decision-making. They have the potential to influence decisions concerning job applications, bail eligibility, credit scores, news feeds, etc. The Report raised auditing algorithmic decision-making for transparency and fairness as a concern. As bias and bias training data sets are routinely introjected into software systems, assigning responsibility for harm caused as a result of discriminatory and inequitable algorithmic decision-making becomes a necessity.
2) Platform Marketplace
Not surprisingly, the roundtable also focused on concerns around privacy, data protection, and consumers rights in the digital marketplace, reflecting the understanding that big data is a core element in a wide range of technologies changing society. The question the roundtable sought to address was whether it is possible to devise more meaningful tools to protect public and personal interests and rights other than the usual online consent tools.
3) Digital Civil Society
Democratic governance entails public participation and informed debate. This requires digital policy literacy for effective reform discussions. The Report has raised the need for tools that enhance participation and evidence-based discussions. It was noted that “technologists and developers are important law reform stakeholders” as they have the potential to analyze and shape access to justice in the 21st century. The recommendation was to enhance digital access to lessen disparities in opportunities and social economic and generational divides to raise individuals’ awareness of their set of digital rights.
4) Regulatory Approaches to Technology
With personal information being used beyond the control and knowledge of most people, the Report highlights that legal innovation and creative thinking are important to address the “law lag”, creating practical and effective solutions beyond the binary “do / do not regulate”. Shifting the focus of privacy laws from individual consent to collective consumer rights and incorporating Ontario’s Human Rights Code protections from discrimination directly into an algorithm were measures discussed to ensure structural compliance. The model of “click-consent” was questioned as providing no meaningful free and informed consent as “the only real choice is to opt-in or miss-out”. Other discussed approaches included consumer rights standardized terms and conditions, the creation of a “digital bill of rights”, smart contracts that allow for line-item consent and mandating certain protections and rights across an array of products and platforms.
5) Digital justice and Equity Principles
An important theme in the Report was technology as the new frontier for access to justice. The Report remarks that inclusive AI is an equity principle which is tied up with fundamental civil rights and due process. The questions on this front which the LCO and Mozilla seek to answer include: “How do we know when an algorithm has been involved in deciding a job application, eligibility for social entitlements, or is adjudicating your legal claim to housing or immigration status? How can you protect reputation online? How are digital platforms, like social media and commerce sites, subject to collective consumer rights, and what dispute resolution mechanisms are needed?” A proposed solution to address these questions involved the creation of a digital rights agenda setting out a framework for: “digital inclusion and access”, “digital bill of rights”, “rights in a smart city”, “digital due process”, “regulatory sandboxing”, “professional regulation and practices”, “social scoring and algorithmic black boxing”, “digital democracy”, and “new frameworks for informed online consent”.
6) Future of work and employment relationships in the gig economy
The Report flags an increasing proportion of Ontario’s workforce earning income through short-term and precarious employment contractual conditions, which means “more people working where the law may be unclear.” According to the Report, key issues to be addressed on this front include: greater clarity for the employee / contractor relationship; how labour and employment rights apply to digital platforms; whether “terms of service” in contract relationships are based on free and informed consent; the fairness, transparency, and protection from discrimination in assigning work via algorithm; dispute resolution and mandatory arbitration; and whether there is a “right to access” work only available through private digital platforms.
Stakeholders may want to watch out for new developments in the digitalization process if they do not wish to miss out on meaningful opportunities to participate in important conversations that are shaping the regulatory landscape that will impact AI, big data and technology generally.