Digital Content Governance and Data Trusts — Diversity of content in the digital age

February 2020

Digital Public

On this page

List of acronyms and abbreviations

CCFA
Cape Cod Fisherman’s Alliance
PIPEDA
Personal Information Protection and Electronic Documents Act
UNESCO
The United Nations Educational, Scientific and Cultural Organization
USMCA
United States-Mexico-Canada Agreement on Trade

Disclaimer

This document has been prepared for the Department of Canadian Heritage and the Canadian Commission for UNESCO by Digital Public. The views, opinions and recommendations expressed in this report are those of the author and do not necessarily reflect the official policy or position of the Government of Canada. Responsibility for any errors, interpretations or omissions lies solely with the author.

Introduction: A practical approach to digital content governance

The Department of Canadian Heritage is committed to taking a proactive approach to ensuring the diversity of content in the digital age. As part of its international engagement strategy, Canadian Heritage partnered with the Canadian Commission for UNESCO in 2019 to assemble a meeting of international experts on achieving several core policy goals, toward preserving the character and availability of diverse content in digital media ecosystems. The output of that meeting focused on three priority considerationsFootnote 1:

The meeting also highlighted the potential to use digital and data trusts as a set of tools to begin experimenting with models to achieve those goals. This report is an extension of that recommendation, and explores, in-depth, the potential applications of data trusts to achieve those goals in the digital content ecosystems.

This report examines the role of data trusts by situating them in digital content and distribution markets. This paper begins by explaining the types of digital content markets it’s concerned with, then moves into the mechanics of data trusts and the role of data governance, and then explores specific points of intervention in digital content production and distribution markets. The analysis concludes with several recommendations for ways to experiment with data trust-based approaches to achieving the Department of Canadian Heritage in the context of its engagement strategy on Diversity of Content.

Digital content

One of the primary challenges for any approach to regulating digital content is that it is definitionally ambiguous – and can refer to an almost infinite number of things. Digital content can refer to the digital representation of anything, from video games, to police camera footage, to purchasing histories, to your health prognosis. At times, the ambiguities that digitization can create are used to distract or hinder public attempts to regulate digital ecosystems – and, rightfully so – overbroad regulation can also be damaging. While this analysis focuses on cultural and news-focused digital content, the ambiguity in the terms themselves can make broad policies unwieldy.

Rather than try to capture the entirety of digital content, this report focuses on market structures and reasonable stakeholder expectations to frame points of intervention. Digital content markets break into three, primary categories: production, distribution, and consumption. In each of these market stages, stakeholders hold important rights and powers.

The primary value of data trusts, as an operational tool, is to consolidate the leverage afforded by collective rights to make digital marketplaces more transparent, diverse, and equitable.

Introduction to data trusts

Digital transformations are re-orienting the relationship between residents, governments, and industry. Those changes occur, for the most part, without the kind of public transparency, participation, or accountability necessary to earn trust and legitimacy. And while both governments and industry players are going to great lengths to conduct consultations, publish strategic and ethical commitments, and call for greater regulation – it can still be prohibitively complicated to understand, let alone pursue, your rights in digital ecosystems. Here, we use the term right to mean an interest recognized by law - those interests can be diverse, but the underlying idea is that a 'rightsholder' is a person or organization whose interests and claims the law will protect in a particular context.

That’s particularly true in digital content systems, where a host of new tools, platforms, infrastructures, billing arrangements, and ways of measuring behavior have fundamentally changed the way in which media markets operate. For many stakeholders in digital content markets, there aren’t accessible means of understanding the determinants of their own success. In order for digital content markets to sustainably function, let alone meet public policy requirements around diversity, equity, and political speech, there’s a clear need for structures to broker the agreement between public interests and private industry.

Data trusts are a way to build transparent, accountable governance structures over digital property and rights.Footnote 2 Trusts are legal instruments that assign management authority over an asset to a steward, a trustee, who uses that asset to achieve a purpose on behalf of the ultimate owner of that property, the beneficiary (or group of beneficiaries). Data trusts take that same legal architecture, but apply them to data, and the rights that digital transactions implicate, to achieve the purpose for the beneficiary. Civic data trusts adapt traditional trust structures, creating governance models – like town councils – to manage the digital rights instead of appointing a single trustee. Trusts are common law instruments, and vary both by national implementation and by theory of law. While a number of legal systems don’t formally recognize ‘trusts’, a significant number of countries have similar ways to legally delegate authority. Other jurisdictions use the term ‘trust’ to refer to multiple things, including charitable organizations. In Canada, data trusts have become a part of the proposed PIPEDA reform and are broadly recognized as a way to create fiduciary accountability in data sharing relationships.

Trusts are different than other assignments of authority in three, crucial ways: (1) they are broad grants of authority, as opposed to contracts, which are specific; (2) they create what are called fiduciary duties – essentially, legally enforceable requirements that the trustees use the asset in the beneficiary’s best interests, selflessly; and (3) they are rooted in the underlying ownership of the data and digital rights themselves, and so survive transfer. In other words, there are a range of legal characteristics of trusts that make them a particularly robust toolset to manage digital rights negotiations – especially when those negotiations have to balance the interests of multiple, at times competing, parties.

Elements of a trust

Grantor/Settlor:
Person or organization that owns the asset and gives it to the beneficiary (trust)

Asset:
The data or digital right to be managed

Trustee:
The person, organization, or governance body that oversees the asset

Purpose:
The thing the trust is designed to achieve

Beneficiary:
The person or group of people the trustee serves, and the owner(s) of the asset

Roles for data trusts

Ultimately, of course, the structures of trusts are only interesting if they’re specifically useful to achieving a set of objectives. While data trusts are a nascent toolset relative to their potential, there are already a number of data trusts in existence demonstrating some of their highest-potential and highest-demand use cases:

Professionalizing governance

Digital markets and supply chains are built on mutual dependence – often in informal or inexplicit arrangements. Data trusts are a useful way to mutually enforce multi-party data sharing relationships that create benefits for the broader ecosystem. Some examples of this in content governance might be agreeing to common data licenses, shared transparency requirements around the role of algorithms in distributing content, and/or sharing metadata about distribution and consumption patterns.

Sovrin is a digital identity company, who aspires to become a global public utility. In an effort to build a multi party marketplace with credible standards and limitations around data sharing, Sovrin based its governance on the Sovrin Trust Framework (PDF Version, 423 KB). The Trust Framework sets out market principles, standards, adaptation mechanisms, and data sharing relationships. Others: The Scott Family Trust (owns The Guardian)

Cape Cod Fisherman’s Alliance (CCFA) is a network of fisherman policy advocates and researchers concerned about ensuring that the digitization of their industry doesn’t leave them further behind. With significant federal reporting public oversight and economic capture concerns the CCFA is building a data trust to leverage its role in the digital supply chain to protect its members’ practical and financial interests. Others: Data for Black Lives the Light Collective

Collective action + bargaining

As with traditional labor organizing, digital markets create significant asymmetries that often require collective action to rebalance. Data trusts can be a vehicle for digital rights holders to consolidate the leverage created by those rights, and use them to negotiate for better terms. Other times, data trusts can be vehicles for powerful actors to consolidate their claim in digital spaces – in the United States, both Republican and Democratic party operatives are using data trusts (not always the legal tool) to broker access to their voter files. Some examples of this in digital content governance are content production collaboratives, civil rights and political advocacy groups, and – to an extent – regulators setting standards on issues like disinformation, privacy, and data protection.

Research

In a range of industries, including digital platform regulation itself, there are large inefficiencies between data production and use. That’s particularly true for researchers, who play a critical role in driving public and policymaker awareness about often overlooked and unanticipated aspects of digital ecosystems. One prominent use of data trusts is pooling information between historically disparate sets of actors, for the limited purposes of public interest research. An example of this for the digital content space might include social platforms, university researchers, civil society groups, and end users agreeing to a process to collaboratively share data about political misinformation.

Cincinnati Children’s Hospital Medical Center works with a network of pediatric research and care networks all of whom share data toward advancing their complementary medical research. They are building a data trust to formalize and standardize the sharing relationships between the nodes in their network as it grows and produces commercially valuable outputs. Others: Johns Hopkins University Hospital Data Trust the Structural Genomics Consortium Open Science Trust

Compliance

One of the major functions of any regulated industry is reporting to authorities – and many of those reporting processes are becoming increasingly digital. The digitization of compliance reporting has had both benefits – there are a number of companies that offer data protection and privacy compliance tools – and costs – the privatization of that function creates a potentially asymmetric market itself. For digital content markets, the levers of compliance enforcement and regulatory authority vary by jurisdiction, mandate, and relationship to industry, but there is a relatively uniform increase in the formalization of data regulation, as well as the emerging markets based in the digital transformation of pre-existing compliance workflows.

BrightHive, Sightline Innovation, and Truata are all companies that use the term data trust as part of offering commercial services that variably involve legal data trusts. BrightHive and Sightline both offer digital platforms that enable users to manage data access and data management compliance reporting with multi party governance. Truata is a third party data anonymization service that ensures data transfers are compliant with prevailing compliance standards.

Oversight + safety

Social media and digital content platforms have to mediate between competing interests – often in ways that benefit from engagement with users or other stakeholders. Platforms increasingly use third-party civil society organizations to either set socially impactful standards – for example, around news source credibility – as well as working with independent oversight groups. One compelling use of data trusts is to create a credibly independent and legally accountable oversight body, toward ensuring the best use of the underlying platform to achieve public interest goals.

Ultimately, the value of using data trusts in digital content ecosystems is not that they’ll create new forms of authority – it’s that they’ll enable existing, but disparate, stakeholders to engage in collective bargaining, appoint independent governance, and create a means for meaningful participation in the way norms and policies are made in digital ecosystems. Data trusts are a tool for formalizing markets, supply chains, and governance mechanisms in ways that keep everyone involved accountable to public expectations, in order to legitimize digital transformations. Said simply, data trusts are a way to embed formal data and digital content governance.

Facebook’s Oversight Board is organized as a data trust in the United States. As it’s currently created, it’s administered by the former head of Article 19 – a digital human rights non profit organization – and has the ability to set and change Facebook editorial policy. While the Oversight Board’s authority is less traditionally rooted in “data,” – its supervisory authority is binding to Facebook’s corporate governance and is inevitably executed through digital investigation and means. While it’s too early to know how effective the Oversight Board will be, its selection by companies like Alphabet and Facebook, as well as governments like the UK and Canada, demonstrate their potential to bridge public private policy divides.

Digital content governance

Digital content governance is the act of defining the terms and characteristics of an information ecosystem, including access, types of participation, remuneration, amplification, safety, revenue generation, user roles and control, among many others. Digital content governance decisions are often made across multiple parts of a media or platform ecosystem – from business strategy in the boardroom to the algorithmic indicators used to amplify particular pieces of content in an engineer’s office to the content producers and retailers who rely on platforms for their livelihood. Yet it’s exactly these decisions that define the equity, transparency, diversity, and integrity of digital content ecosystems. The best ways to ensure that digital content governance systems result in public interest outcomes is to formalize their operation, and create clear expectations around participation and dispute resolution.

Many participants in Canadian Heritage’s engagement strategy on Diversity of Content in the Digital Age recognize interventions in digital content governance as a strategic priority to achieve increases in the transparency and diversity of digital content ecosystems. They also recognize that designing and imposing any layer of governance adds non-trivial complexity. Any meaningful governance mechanism, in order to be effective, will need to balance: independent credibility from each stakeholder group, sustainable revenue to support operations, and, at a minimum, some form of meaningful leverage over extra-national actors.

Goals of Canadian initiative on diversity of content in the digital age

  1. Access and Discoverability
  2. Remuneration and Value Distribution
  3. Diversity and Quality of Information/Content
  4. Transparency and Algorithmic Accountability

Practical considerations

Each of these considerations is a material element of any credible digital content governance mechanism – but they will necessarily vary by context. Digital content platforms have internationalized and vertically integrated to the point where they are no longer discretely recognizable as one type of company with a traditional precedent. In the same way that it’s nearly impossible to create a bright line definition for ‘digital content’, it’s equally difficult to create a bright line definition for ‘digital content governance.’

Rather than focus on the entities that may play a role in digital content governance, this analysis focuses on typical roles in content governance and distribution to identify rights holders, as a way to identify points of intervention for collective action and governance.

Digital content markets and governance

There are a range of ways to influence digital content markets, especially when including the broader palette of tools available to government actors. Data trusts, while capable of administering delegated authorities from governments, don’t have any inherently regulatory character. This analysis is explicitly focused on applications of data trusts to digital content ecosystem governance that doesn’t rely on sovereign regulatory authority, and instead draws on the potential for collective organization of supply chain stakeholders. Similarly, there are a range of approaches to managing digital content governance by manipulating the companies that own platforms through tactics like shareholder activism, industrial self-regulation, and degrees of nationalization – all of which could use trusts, but aren’t specific to digital content markets.

Data trusts as a model of intervention identify opportunities for mutual and public interest, embedding governance in the functional management or oversight of the core workflows and transactions in a marketplace. Content markets are an existing set of structures, standards, and authorities that are being reorganized through digital transformation – and can be renegotiated through a range of approaches to collective governance. This analysis breaks digital content markets into three broad roles: (1) production; (2) distribution; and (3) consumption. This categorization is broad, focusing on the essential requirements of digital content markets.

Digital content markets, however, are also complicated by third-party interests, like advertisers. The escalating efficiency of digital advertisers attracts surveillance customers from a broad range of interests, increasing the social and political complexity of governing their services. There’s an inherent tension between the integrity of content marketplaces and the role of advertisers, who often represent their primary source of revenue, that predates the Internet. There are ways to use data trusts to create credible public transparency, accountability, and governance for advertising and content distribution – but they also aren’t necessarily specific to digital content markets.

This analysis focuses on direct content markets, which also tend to be more analogous to digital transformation in adjacent industries, and the models of collective action and leverage different stakeholder groups might use to compel diversity of content, transparency, and fair markets.

Models of influence in content production

Digital content markets begin with content producers. For digital content producers, digital platforms are significant parts of the full life cycle of the commercialization process – blending the means of production with the tools of distribution and consumption. In many cases, blended platforms also make a claim to ownership of content produced using their tools – either in whole, or with licensing that grants broad rights. Here, content producers suffer from both the asymmetrical nature of the negotiation, as well as from being isolated as individual producers without much collective representation. Beyond labor and/or commercial negotiations, professional associations and industry bodies are often important forces for defining socially and politically contextual norms, like dimensions of diversity in content.

Arguably, the largest ‘innovation’ in digital content production was the move toward behavioral surveillance and targeting. Analogue media, while financially dependent on advertising, wasn’t able to attach influence to content with anywhere near the granularity or breadth that digital platforms are capable of delivering. As a result, we have more potential to understand the value and impact of content, participate in the design of the content we want amplified, and shape the way we discover new things. The barrier, in many instances, is that information about the functioning of the digital content marketplace is protected by trade secrets, as are the means of distribution. At the same time, marketplaces are typically regulated – not only based on outputs, but also the rights and dependencies of participants. Those rights protect market participants and creators, and both offer important entry points for increased transparency, diversity, and using data trusts to improve governance.

Table 1: Use of Data Trusts in Content Production
Stakeholder Right Legal Right Data Trust Use Case Example
Creation Copyright/Licensing Content producer collaboratives to negotiate for market disclosures and collective representation, participation, and remuneration rights. Canada’s Innovation Asset Collective, content producers could use data trusts to pool asset ownership and access for collective negotiating power.
Likeness Representation Using collective action tools to govern the ways that digital representations and reputations are used to shape content distribution markets. National Football League Players Union: Data sources could pool limited representation rights to a data trust to broker legitimate data usage and remuneration.
Market Entrant Contract Like industry associations, balance investments in market, content and vendor quality with negotiating for improved transparency, equity, and accountability. Alliance of Canadian Cinema, Television, and Radio Artists: Data trusts could set terms for digital content production indexing, diversity, and transparency standards.

Models of influence in content distribution

The most powerful gatekeepers in digital content markets are the distribution platforms, which is also true in many analogue content ecosystems. Distribution platforms – whether television channels, newspaper layouts, or algorithmic newsfeeds – have always made their business at the nexus point between content production, consumption, and revenue. Every content distribution platform performs a gatekeeping function, choosing which content to distribute, amplify, and monetize. That gatekeeping function is inherently political and, historically, subject to significant public scrutiny, regulation, and criticism. Nearly every content distribution organization struggles to maintain the editorial integrity that defines its brand, while competing for advertiser revenue, and ensuring a valuable product for their consumers.

Digital content production platforms initially evaded a significant amount of the liability for their gatekeeping through a combination of focusing on the responsibility of end-users, automating content moderation, and referring disputes to customer service or arbitration, all of which are opaque processes. That approach is starting to fray, however, as digital content distribution platforms begin to meaningfully grapple with the social complexity and importance of their work. Large, vertically integrated digital content distribution platforms are increasingly recognized for playing a role in major scandals, whether contributing to genocide, manipulating democratic elections, or exacerbating the spread of pandemics. As a result, there’s an unprecedented opportunity to structure participatory governance in the most critical layers of digital content distribution.

There are a range of applications for using data trusts as a vehicle for building more participatory and inclusive governance in key functions like: overseeing the design and adaptation of content amplification algorithms; articulating, localizing, and enforcing political speech policies – including for advertising targeting; and representing multi-stakeholder interests in the formulation of content licenses and redistribution limitations. While none of these functions are simple, they are all critical policy challenges where digital content distribution platforms have expressed willingness to engage – especially where it mitigates their representational or financial liability.

Table 2: Use of Data Trusts in Content Distribution
Stakeholder Right Legal Right/Liability Data Trust Use Case Example
Content Amplification + Moderation

Free speech

Bias, Unfair commercial practice, Unjust inurement, Incitement

Content distributors could use data trusts to create bounded design and governance networks to oversee, publicly review, and adjust content amplification systems to protect the interest of the public, as well as marginalized groups. A broad range of predominantly local newspapers use reader advisory boards to evaluate editorial decision-making, suggest adjustments in coverage, and feedback on customer disputes.
User Profiling

Know Your Customer

Breach, Negligence

In order to achieve diversity of representation, user profiling functions and metrics could be reviewed and independently overseen by a data trust with evolving governance. Facebook’s new Oversight Board is organized as a data trust in Delaware, and has a broad remit to protect consumer interest and set Facebook policy.
Data Brokerage Contract Data trusts could be used to provide independent oversight of the data use and brokerage of digital content platforms, managing regionalized compliance and consumer protection. The Silicon Valley Regional Data Trust oversees the management and use of a database of student information, toward ensuring the quality and type of data use while limiting student exposure.

Models of influence in content consumption

Although digital content distribution platforms are often thought of as the “top” of the commercial food chain, the primary reason is their ability to centralize attention and demand. By doing so, they have historically been able to isolate and marginalize the role of consumer-driven advocacy tactics, like boycotts. For most digital content platforms, consumption is so globally diffuse that it can be impossible to organize unified advocacy actions and even broad efforts at coalitions have struggled to gain audience.

The emergence of significant political fissures, and regulatory responses around issues like misinformation, have led major platforms to begin to experiment with participatory content governance. There are a number of models for that participation, from assigning moderator authority to volunteers – like Wikipedia and Reddit, to outsourcing political decisions to professional standards groups or content moderation services – like the International Fact Checking Network and Cognizant. As these models evolve, they could grow to include contextualized, adaptable governance – both as independent organizational structures and, potentially, as embedded options in the way that people use digital content platforms.

Data trusts are already used as ways to certify characteristics of digital content products – like anonymization or valid consent. Digital content consumers could also use data trusts to do things like independently report on the degree of diverse content a platform produces, the way that producers are (or aren’t) being remunerated for content, and/or the amount of controversy and user participation in making governance decisions.

Table 3: Use of Data Trusts in Content Consumption
Stakeholder Right Legal Right/Liability Data Trust Use Case Example
Consumer Consumer Protection + Supply Chain Transparency Content consumers could organize data trusts that curate ethical, diverse, and equitable content production and distribution systems. Similar to content ratings or agricultural production certifications. Both Meedan and Stanford’s Internet Observatory are compiling databases focused on misinformation, whose goal is to involve the public in identifying, annotating, and ensuring the integrity of information on digital content platforms.
Data Subject Privacy + Representation Data trusts can be used as collective action tools to pool civil and commercial rights, pursuing a range of legal and creative remedies to exercising their beneficiaries’ interests. The first ever reference to data trusts theorized using them for collective privacy and data rights enforcement.

Isolating the role of algorithms

One of the major distinctions between digital content governance and analogue content governance is that technology platforms have historically abstracted the political dimension of their editorial responsibilities into a combination of user responsibility and algorithms. That policy, in the United States, is best exemplified by Section 230 of the Communications Decency Act – which gives platforms broad protections for the content shared on its platform – which has also been negotiated into a range of bi-lateral and multi-lateral trade agreements like the US-Mexico-Canada Agreement (USMCA). That position is under political contest – the European Union’s recent Copyright Directive, for example, creates liability for platforms trading in protected content.

Data trusts are a useful tool for creating bounded, public interest data sharing relationships between stakeholders in digital content ecosystems. As is true of any maturing industry, both established and emerging actors are beginning to invest in the long-term quality, and thus sustainability, of their field. As the digital transformation and governance functions of content ecosystems mature, they could use data trusts as a vehicle for fiduciary, independent validation of the design of content amplification algorithms, testing and warranties against common failures, or even providing direct avenues of accountability or dispute resolution for algorithmically made decisions. Ultimately, each of these functions is critical – but dependent on an avenue for governance intervention and/or meaningful leverage – both of which can be invisible or impossible to isolate inside of complex content ecosystems. There are promising initial signs in the form of research-focused, public-private data sharing relationships like Social Science Research One – but we’re still in the early stages of understanding the governance implications of those relationships.

Conclusion + recommendations

Digital content governance is one of the most complex and urgent issues of our era – and the way we organize, aggregate, and exert our collective agency will continue to play a critical role in defining our both our present and future. Thankfully, collective action and governance challenges are not new to the human experience – and there are a broad range of tools, interventions, sources of leverage, and promising approaches embedded our laws, markets, and political infrastructure. Data trusts are one of those tools – and they offer an early opportunity to begin experimenting with collective governance structures capable of delivering digital content ecosystems and digital transformations in the public interest.

In order to realize that potential, however, this analysis focuses on five areas in need of ongoing critical research exploration and policy development:

Data trusts are a promising approach to building equitable governance of digital content platforms. They will not, however, realize that promise simply by existing. In order to build a diverse, inclusive, and transparent digital content ecosystem, it will need a wide range of governance systems, tailored to the context in which they operate. By investing in an enabling environment that clearly articulates digital rights, makes it easy to assign rights to stewards, and incentivizes markets around quality and integrity, public interest stakeholders will begin to build the collective action infrastructure and market incentives for a diverse, transparent, and fair digital content ecosystem.

Page details

Date modified: