This article is part of the GeoTech Center’s ongoing work on promoting and piloting data trusts and data co-ops. For more of the Center’s work regarding the Future of Data, check out its related content at the end of this article.
Past Atlantic Council articles have examined how data trusts could be applied to the COVID-19 pandemic and other aspects of the global recovery. This article explores how private and public sector leaders should develop strategies for leveraging data trusts in order to produce new economic value, accelerate innovation, level the playing field of the emerging AI economy, and enable inclusive, win-win economic outcomes for all.
Artificial Intelligence is forecasted to add at least $13 trillion to the global GDP by 2030, a huge and much-needed boost to the global economy, especially given the hit to growth caused by to the pandemic. Nevertheless, there are at least three big challenges that must be tackled as big data and smart algorithms transform economies: the need to level the field regarding access to high quality data, the ability to remain competitive in the emerging geopolitical order, and the broader impact of AI on the future of work. First, data are at the heart of the AI revolution; however, not all actors in the economy have equal access to high volumes of good quality data, resulting in a widening chasm between data-rich and data-poor organizations. Second, liberal democracies must uphold their core values of protecting citizen privacy, which puts them at a disadvantage to their authoritarian competitors that collect, access, and use sensitive personal data for AI development without any legal or ethical barriers. Last, by automating cognitive job tasks AI systems threaten the future of work and call for a radical reinvention of the current social contract so we are prepared for a future where work and income will be intermittent and uncertain for most citizens. Let’s take these three challenges in turn, and see how data trusts can help us find new solutions.
Resolving the privacy vs utility dilemma
Companies that aspire to become winners in the AI economy are constantly struggling to gain access to a wide variety of high-quality data in a compliant, ethical, and secure way. Moreover, there are many barriers to data sharing between organizations. At the core of all these issues lies the “privacy versus utility dilemma.” There is enormous value in accessing and combining raw personal data, but at the same time the privacy of that data must be preserved. If a company collects such personal data it is near impossible to share that data with another company without breaking the law or incurring high compliance costs.
But what if there was an independent organization—called a “data trust”—that collected personal data and had fiduciary responsibility towards the data providers for governing their data’s proper use? Fiduciary responsibility matters because it obliges the data administrators to prioritize the rights and benefits of the data providers over those of the data consumers. Data trusts can ensure the ethical and compliant governance of data privacy (e.g. by ensuring user consent, removing data bias, and de-identifying the data). Data consumers, such as private business or governments, can then contract access and use of the data in a data trust, and focus on the utility. Moreover, citizens—as data providers—could have a direct say in how the data trust and their data are governed. They could elect the trustees, audit their performance, and vote for the organization’s constitution and policies.
Accelerating AI innovation in the new world order
By solving the privacy versus utility dilemma data trusts can also act as brokers for sharing data among organizations. “Minimal Viable Coalitions” (MVCs) of data consumers and data providers can set up a data trust to govern a secure data exchange where they may trade in data, as well as combine data in order to develop and accelerate adoption of innovative AI solutions. Think of, for example, a data trust brokering secure and ethical access to patient data, genetics data, and health insurance data, to companies that want to combine such data in order to innovate new products and services in healthcare, wellbeing, and insurance. Or imagine the potential for innovation around data from smart cities in liberal democracies where fears around citizen surveillance are currently preventing data access, sharing, and collaboration.
In the well-known case of Toronto many people and civil groups reacted negatively when Sidewalk Labs, a Google company, attempted to control the collection as well as the processing of personal citizen data from sensors and cameras. By focusing too much on technology and ignoring sensitive issues around data collection and processing, Google failed to convince Torontonians of its good intentions and had to pull out of the project in May 2020. Perhaps a data trust approach whereby the data trust was run independently and democratically by citizen representatives, could have solved the problem around data privacy and control, allowing Google to proceed with their plans, which included significant contributions to Toronto’s economic and technological prowess, not to mention 44,000 new jobs. China is currently home to half of the world’s smart cities. Liberal democracies will only start to catch up once they start adopting data trusts that can resolve society’s fears around state or corporate surveillance.
Monetizing citizen data to fund UBI
Beyond ensuring ethical use of data and accelerating the AI economy in liberal democracies, there may be an even greater prize for rethinking the governance of personal data through Data Trusts. As our economies dramatically transform due to AI algorithms and data, “business as usual” scenarios whereby the winner takes all are likely to exacerbate income and wealth inequalities. Such inequalities will be compounded by work automation due to AI, which may render millions of people permanently unemployed or underemployed. It is doubtful that liberal democracies can survive the double whammy of huge economic inequality and social injustice. Data trusts could provide at least part of the solution for greater economic equality in an AI economy powered by data by capturing some of the economic value and redistributing it back to the data providers. Think, for example, of a data trust that represents patient data, or the data collected from citizens in a smart city, or the behavioural and consumer data of gig workers. Such a data trust would be able to monetize that data for the benefit of its data providers. Data monetization can be a very valuable business. Recently, American Airlines tapped into a government loan of $4.7 billion by putting up as collateral its loyalty program, effectively a database. Valuations of that database ranged from $18 billion to $30 billion. A data rust that managed such a valuable asset could provide sizable annual dividends to its data providers. As policy makers and politicians are thinking of ways to finance a universal basic income for citizens whose work and income will be affected negatively by automation, data trusts can provide a powerful financial instrument for funding such schemes without resorting to increases in taxation or borrowing and without making citizens more dependent on a welfare state.
What leaders in the private and public sector can do
Explore data sharing use cases: create Minimal Viable Coalitions (MVC) with business and government partners where data is shared in order to solve a problem with significant commercial and social returns; explore legal barriers for sharing data, use data de-identification and federated learning technologies, include the goals of inclusion and ethics in the constitution of the MVC.
Establish data trust cooperatives: professional associations, mutual-ownership companies, trade unions, and smart cities are perfectly placed to set up and run data trust cooperatives where the data of their members can be collected, stored, and monetized.
Enshrine data property rights in law: we need to start thinking data property rights as a fundamental building block of the AI economy. Given that data becomes valuable when it combines with other data, “private ownership” may suggest mutual ownership rather than individual ownership.
Create a regulatory framework for data-as-an-asset: develop the necessary regulatory framework to allow trading of datasets as tangible assets. Such a framework will enable much-needed innovation around new digital financial instruments for funding the creation of data trusts and data cooperatives. Explore distributed ledger technologies as the means to tokenize mutual ownership of a data trust and create new classes of digital assets for data trusts that can be traded in regulated exchanges.
Championing new technologies and data to benefit people, prosperity, and peace.