In short time, the Canadian government has taken important steps toward better AI governance. This progress was evident at Student Symposium on AI and Human Rights held in April by Global Affairs Canada and the Canadian Institute for Advanced Research. Participants heard first hand how the Canadian government is responding to AI’s impact on human rights specifically related to equality, privacy, accountability and freedom of expression. With these issues in mind, Canada still has much work to clarify the regulatory environment in place for artificial intelligence. Below are some observations from the event.

Government as a leader in inclusion

The tech sector has been long criticized for being too white and too male, leading to technologies – and policies – that do not accurately represent the experiences of women and POC. The risk is that homogeneous design times will not properly anticipate the real world applications of technologies, creating AI that reflects the biases of the teams that worked on them.

By contrast, the GAC symposium featured a mix of students from different disciplines and backgrounds, young men and women in equal measure hailing mostly from Quebec and Ontario. I enrolled my own graduate students from Concordia’s Media Studies program to participate in the symposium. Their research led to presentations from two groups concerned with the implications of AI on the Canadian labour market as well as the ways that automation might increase race and gender discrimination.

The Symposium exemplifies how the government can lead by creating inclusive spaces. Clearly the government might not be a leader in all cases, especially involving the relationship between AI and indigenous peoples, but the support given to allow students to participate in this conference clearly indicates a commitment to public participation that will need to continue in order to have democratic oversight of this new technology.

A New Generation of AI Scholars?

When we started preparing for the symposium, my students, mostly women, were concerned that they knew nothing about AI governance – a reminder that the rhetoric of high-technology disfranchises many important political discussions. Siva Vaidhyanathan cautions that presumptions like these, about who interacts with technologies and who can claim expertise, means “we forge policies and design systems and devices that match those presumptions”. Overcoming their concerns about who can be considered an ‘expert’.

Students quickly learned the key issues in AI governance. One group considered the ways that AI-led systems for governing divorce proceedings might exacerbate the inherent factors tied to gender inequality between couples seeking a divorce.

Many students raised concerns that Canada lacks accurate good public knowledge about artificial intelligence. Instead, discussions of AI in Canada have tended to use the language from Silicon Valley. Disruption. A fourth industrial revolution. In reality, the hype around AI can foreclose a good policy debate, making AI seem like a complete disruption rather than part of a much longer pattern of technology in society.

That these students could go from being concerned about their lack of expertise to knowledgeable participants at the forefront of AI governance is a potent reminder that the lack of inclusion seen in the conversations around AI seems more to be a lack of engagement than ability.

The capacity of these students to learn and advise on AI supports calls for new researchers trained in assessing the impacts of algorithms. Understanding how AI will impact society requires interdisciplinary research, especially for the social sciences and humanities to understand its lived impacts and our everyday understandings of new technology.

Concentration in the AI Industry

AI does not build itself in a vacuum, nor do our understandings of AI exist without inaccuracies and bias. What’s missing remains an acknowledgement that AI is the key research area of major platform-based companies such as Google or Amazon. These companies face inadequate public scrutiny in spite of a long history of mistakes and abuses in developing their tech. Good AI governance cannot repeat the past missteps of platform policy.

The growing concentration of power amongst the leaders in AI requires us to be aware that the future of AI is political as much as policy-based.The willingness of the government to partner and promote its industry contacts could likewise impede its ability to hold these firms accountable.

Questioning the Language of Disruption

Student presentations questioned how we talk about, describe and imagine the impacts of AI. These representations, as media students students know, are powerful. There is already a concern that too much AI discussion focuses on the omnipotent AI to come ignores the rather mundane applications already here. My students found that much of the hype around AI sidesteps holding it accountable to already existing policy frameworks.

AI policy tends to try and develop new paradigms rather than repurpose existing frameworks. Treating AI as an unprecedented novelty impedes the government’s ability to create clear guidance about AI development and implementation. Why do we need new guidelines for Ethical AI when we have the Canadian Charter of Rights and Freedoms? Canada could easily advances its approach to AI governance by interpreting the charter for AI research and development.

The innovation agenda, to borrow a popular government phrase, cannot continue to ignore the crucial and very tangible step of interpreting the design and deployment of machine learning systems through effective and already existing humans rights frameworks and labour law.

Global AI Governance, Locally

The progress made by key government agencies suggest that these issues will be addressed soon, but for now there remains a global gap in AI leadership. Considering the impact of General Data Protection Regulation globally, Canada could be a global leader if it adopts strong national initiatives.

The rush to consider the GDPR as a global data framework also indicates there is a first mover advantage in developing new digital policy. Modifying existing frameworks to suit AI, and considering the experiences of a diverse population are important to building AI policies that serve all Canadians effectively.

Canada can can be a global AI leader if we continue at this pace to implement humane, just and inclusive AI governance.

 

I. Comments

1. Who we are: We are submitting this intervention regarding the Part 1 Application to disable on-line access to piracy sites submitted by FariPlay Canada as individuals. We are researchers with expertise in the analysis of telecommunications policy, Internet infrastructure development, and the implications of network control technologies that shape Internet access and use. This research has been supported, in part, by the Social Sciences and Humanities Research Council of Canada (SSHRC). We are grateful to Jean-François Mezei for collecting the data used in this analysis. The views expressed herein are ours alone and should not be interpreted to reflect those of any affiliated organizations.

 

2. Technological reality: As researchers that both produce and consume copyrighted scholarly works, we recognize that copyright infringement can have negative revenue implications for rightsholders. However, the history of the development of the Internet over the past two decades is replete with examples in which attempts by Internet service providers to use blacklists have proven to be wholly ineffective in achieving their objectives as people learn to use new technologies and obfuscation techniques to bypass such blocking mechanisms. The FairPlay Application provides little credible evidence that the type of approach it is proposing is likely to have any meaningful impact in further reducing “piracy” by the 7% of Canadian households (i.e. 2-3% of population at 2.5 persons per household, if we were to assume there is only one pirate in each household, which may or may not be a valid assumption) FariPlay contends are engaged in unlawful copyright infringement. We suspect sophisticated network engineering and technical teams in fixed and mobile Internet access providers that are part of the FairPlay coalition would agree with our assessment of the likely ineffectiveness of the proposed solution to achieve its stated objective of reducing online copyright infringement.

 

3. Economics of false positives: Furthermore, we are particularly concerned that adoption of the type of approach to addressing copyright infringement FairPlay has proposed will lead to a war of attrition between Internet Piracy Review Agency (IPRA) blocking regime and the so-called “pirates”. Without any substantive reductions in copyright infringement, the emergent war of attrition around IPRA’s blocking regime with the pirates will run a significant risk of restricting the ability of others (i.e. innocent bystanders) to access the wide world of legitimate content and sources of information that are available from open Internet (i.e. false positive errors). Even if we were to accept FairPlay’s unsubstantiated claim that IPRA’s blocking regime may be beneficial in terms of reducing financial losses to rightsholders, these benefits must be viewed in terms of the costs of false positive errors expected to occur if the Commission were to adopt FariPlay’s request. From an economic perspective, the risks of false positive errors are likely to growth with time as IPRA learns that its blacklist is ineffective in reducing copyright infringement and responds by expanding its scope to a broader range of Internet resources to achieve its laudable objective of fighting “piracy”.[1]

 

4. Core legal flaws: In addition to technological and economic flaws in the logic of FairPlay’s proposal outlined above, it has a number of legal flaws that we submit are important to for the Commission to consider:

 

5. (A) Statutory authority: There is no specific statutory basis under which the Commission has been given the authority to address copyright infringement. FairPlay’s various arguments that the Commission has the authority to adopt its proposal are highly contorted and not sufficiently credible to withstand legal scrutiny on appeal. Adopting FairPlay’s proposal would be inconsistent with the scope of the Commission’s specific statutory objectives as allocated by Parliament under S.7 of the Telecommunications Act. We therefore submit there is a strong prima facie case for rejecting the Application due to its highly tangential relationship to the specific statutory basis under which the Commission is empowered to operate (i.e. without devoting any further scarce Commission resources to consider the merits of the FairPlay proposal; issuing requests for information or a notice for further consultation).

 

6. (B) 2006 Policy Direction and Commission practice: Even in various previous proceedings where the Commission has specific statutory authority to regulate activities of telecommunications service providers, over the past few decades it has exercised significant forbearance in exercising its authority so as not to “interfere with market forces” too much (e.g. price/service quality regulation, wholesale access to transport facilities, misleading advertising, consumer protection, etc.). In the hierarchy of public interest needs and Commission’s statutory responsibilities, we submit that it may be more relevant to focus scarce agency resources to achieve its specific statutory objectives in terms of affordability of access to reliable and high-quality communication services than venturing into the realm of copyright protection. The FairPlay Application does not provide any evidence to support its contention that the Canadian judicial system lacks adequate copyright protections necessitating the Commission to substantially interfere with legitimate “market forces” in the development of Canadians access to the open Internet (i.e. per 2006 Policy Direction), with limited expected benefits in terms of reduced “piracy”.

 

7. (C) Limited liability of common carriers/common law: In addition to its basic technical and legal flaws noted above, adoption of an extrajudicial Internet blocking scheme like IPRA is clearly inconsistent with the conceptualization of basic telecommunications service providers as common carriers.[2] This is not surprising since the FairPlay proposal only focuses on the costs of copyright infringement to rightsholders, without considering potential economic harm to third party interests that will arise when legitimate material is caught up in IPRA’s blacklist. The logic of FairPlay’s proposal is fundamentally flawed as it looks only at one type of economic costs (i.e. from “piracy” imposed on rightsholders) and omits those associated with inevitable false positive errors that using a blacklist will entail. In other words, the proposal does not account for the real possibility of “who pays” when something goes seriously wrong with the implementation of the blacklist. It certainly won’t be IPRA, CRTC, or rightsholders who will compensate businesses and individuals adversely impacted by material errors in the composition of IPRA’s blacklist. Potential damages from such errors will therefore be effectively allocated to providers and consumers of lawful content whose legitimate economic interests becomes collateral damage to the ensuing war of attrition between the blacklist administrator IPRA and innovative pirates. When damages from false positive errors are particularly large, business and consumers may start to seek compensatory remedies through the courts from the Internet service providers who had little option but to implement IPRA’s blacklist that contained the relevant error. The potential for uncompensated third-party damages under the IPRA regime can ultimately erode justifications for limitation on civil liability Internet service providers generally enjoy as common carriers acting as neutral conduits for lawful communications.[3]

 

8. Summary: Based on technological, economic, and legal reasons outlined above, we therefore oppose the Application for three key reasons: (a) copyright infringement is not within the scope of the Commission’s statutory authority, (b) the proposed blocking regime will do little to reduce “piracy”, and, (c) it does not account for economic damages caused by false positive errors to third party entities, which can create substantive liability risk for various stakeholders in the broadband ecosystem (e.g. users, technology companies, providers of legitimate content and information, Internet service providers). We submit that it would be a costly error for all stakeholders, including FairPlay members who apparently believe their proposal may actually reduce piracy, if the Commission were to consider this Application any further. If FairPlay members have a serious issue with the competence of Canadian courts and copyright laws, they have the option to engage with elected representative to improve the judicial remedies that are available to them and strengthen Canada’s copyright laws as they consider appropriate.  

 

 

[1] Please note that this economic incentive problem on the part of IPRA will arise as it tries to achieve its objective of reducing piracy; independently of how the proposed entity is funded. Funding the agency with fees from rightsholders will only accentuate this problem, but public funding will not eliminate it. For a general discussion of the interplay between different types of errors in the operation of legal/regulatory mechanisms see Rajabiun, R. (2009). Private enforcement of law. Chapter 2 in Encyclopedia of Law and Economics, Criminal Law and Economics, Garoupa, N. ed. Edward Elgar. https://www.e-elgar.com/shop/criminal-law-and-economics

[2]McKelvey, F. (2017) The Internet Was Always a Common Carrier. http://www.amo-oma.ca/en/2017/09/07/the-internet-was-always-a-common-carrier/

[3] This would be rather ironic for telecommunications service providers who are participants in the FairPlay coalition. We suspect short term incentives to increase revenues that some vertically integrated members of FariPlay hope IPRA will enable might be clouding their long term thinking as it relates to the economic and legal costs that going down this road might have for them.

 

Intervention Regarding Part 1 Application to disable on-line access to piracy sites 8663-A182-201800467