Over the past few months, I’ve briefly mentioned the concept of Workers Data Trusts, and in this article I want to expand on it a bit more and explain why I believe the union movement must urgently grapple with this concept.
What are Data Trusts?
Simply put, a data trust is a trust that exists to steward the collective data of a group of people.
“a legal structure that provides independent, fiduciary stewardship of data” (via Open Data Institute)
It uses the same legal framework as other kinds of trusts — and is legally similar to the trustee structure for industry superannuation funds, where retirement savings of Fund members are held in trust by trustees.
The independence ensures there are no conflicts of interest — i.e. that the trustees operate in the best interests of Fund members. The fiduciary responsibility under law is (one of) the highest level of legal responsibility, and involves “impartiality, prudence, transparency and loyalty” to the trust beneficiaries (the Fund members).
The quote from the ODI links to a longer article about the status of thinking and policy development for the concept of data trusts. It’s a long read, so here’s a key take away about the concept:
- “We understand, however, that others may be seeking to advance the use of data trusts for different purposes. Sylvie Delacroix’s and Neil Lawrence’s proposal for ‘bottom up data trusts’, for example, describes a scenario where ’data subjects choose to pool the rights they have over their personal data within the legal framework of the Trust’. They seek to rebalance the respective control that corporations and individuals have over personal data, and provide a legal mechanism to empower data subjects to choose between different approaches to data stewardship that reflect their preferences and needs. Where data portability and other rights – and services like personal data stores – give people the ability to themselves decide who can access and use data about them, this use of data trusts seeks to enable people to appoint others to make those decisions, based on the argument that individualistic approaches to data stewardship are flawed.”
- “Data trusts also represent a way of stewarding data by and for the benefit of defined communities. In this context, data trusts are being discussed as a means to limit or prohibit access to data, and protect and maintain it as a community asset.”
I previously linked to this Guardian article by Peter Lewis about data trusts – have a quick read if you haven’t already.
Why are data trusts relevant to unions
There is a vast amount of data created about workers. This data is increasingly used by employers and large corporations to control and surveil workers, both at work and outside of work.
As I wrote in an earlier issue of my newsletter, a whole host of data that is created has fundamental consequences for the future of power at work.
Examples, identified by Uniglobal’s Future World of Work program, include:
- Semi-autonomous systems like HR and hiring platforms
- Scheduling tools, that impact on who works, where and when
- Productivity monitoring, including warehouse scanning, call-centre seat monitors, and keystroke monitors
- Fraud detection and facial recognition, including data-mining of social media accounts of employees
Each of these (and more) data collection systems represents intrusion of managerial control over the working lives and dignity of working people.
Uniglobal secretary Phil Jennings (a leader whom I have admired for many years) said back in 2017:
Data collection and artificial intelligence are the next frontier for the labour movement. Just as unions established wage, hour, and safety standards during the Industrial Revolution, it is urgent that we set new benchmarks for the Digital Revolution.
This Uniglobal framework is an excellent primer to assist unions to understand what is at stake for workers’ data and privacy:
Data has been termed the new gold. It is traded, analysed and used in marketing, advertising and human resource management. It is also the building blocks of artificial intelligence and algorithms. By 2030, it is estimated that 15-20% of the world’s combined GDP will be based on data flows. It too is the very foundation of the myriad of new businesses and services that are increasingly individualising many aspects of our economy and society, namely the platforms of the so-called gig economy.
As citizens, we daily leave a data trail behind us: from what we search for on Google, to the apps on our mobile phones, from rides we take in taxis, flats we rent, from what we buy, to our loyalty cards, our health records, phone calls to customer services. Not to mention the places we visit, emails we send, Facebook friends we have and tweets we write. Doing all of this provides companies with data – about us and our network of friends. Data is simply the biggest gift we don’t realise we are giving away.
We also provide data as workers—our CVs, our biometric data such as our fingerprints or iris scans, and the abundant data mined on us as employers monitor our workflows. Data, or rather sets of data from within and outside of the company, are also used by management in human resource decisions. Who gets hired? Who gets promoted? Should someone be fired or cautioned? Are the workers productive today and if not, why not? The application and use in companies has even spurred the question whether data is taking the human out of human resources.
Whilst data protection and privacy laws do exist in various forms in many countries, the data derived from monitoring workers is not specifically covered by these laws.
Unions need to grapple with the creation, collection, analysis, storage and sale of workers’ data by employers and corporations.
How can unions address the data power asymmetry?
Australian law provides few or no protection of workers’ data.
One way to address this is for unions to use the collective bargaining framework and legal process. Again, the Future World of Work project provides a schema for unions to understand and design EBA clauses about workers’ data.
This is vital. The first step for unions to start protecting workers’ data is to develop model clauses, based on the 10 Principles outlined by Uniglobal, to ensure workers, delegates and union representatives have access to and influence over the data collected about them.
All of these lifecycle stages have potential for collective bargaining claims. And the truth is that workers can only truly negotiate collectively in the face of data collection on individuals by corporations.
However, when it comes to worker data (and also their data as individuals, patients citizens and consumers), an enterprise by enterprise approach is not sufficient.
Many of the tech companies that provide data services to employers are massive, venture-capital funded multinational corporations (like Cisco, Microsoft, etc).
The employer may use the database provided by Cisco to manage its HR or workers data, but have almost no control over what Cisco does with it. Employers may just pump workers data into a massive “data warehouse” provided by Amazon or Google, with no consideration to the data policies of that tech provider.
Similarly, some employers are part of “group entities” (or franchises), where the company at the top of the foodchain is not the employer but nonetheless accesses and controls data created by its subordinate entities. The entity exercising that data control may not be the workers’ employer, and so an enterprise bargaining approach will/may not cover the top group entity.
The urgent next step: Workers Data Trusts
Conceptually, imagine that workers’ data is like superannuation. Unions can create a data trust — as a legal entity with trustees — and then negotiate with groups of employers on behalf of members and workers to gain control over the data of the beneficiaries of the trust (the workers).
Why do this?
Firstly, workers need to gain collective control over their own data.
Only unions understand this, and only unions are in a position to organise workers collectively to reassert workers’ control over their data.
The only way to balance the economic and social power of employers and tech giants is through collective institutions organised by unions.
Secondly, this is a familiar and effective model for unions and workers. The closest similar example is superannuation trusts managing money for members — there are also older models like Credit Unions — which are focused on creating a “collective system of rights and accountability, with legal standards upheld by a new class of representatives who act as fiduciaries for their members.” (Link)
From a technological point of view, it is easy for a data trust to hold copies of the data created about or collected from their members. A data trust can then negotiate with employers and corporations on how the trust members’ data is used, analysed and accessed — and can alert members to how corporations use the data.
How the Trusts would work in practice
How would a Workers Data Trust work in practice? Well, there are lots of other kinds of data trusts already operating that give us an indication (most of them private, corporate trusts, but some are non-profit trusts — for example, Facebook has created a private non-charitable trust to manage some of the data it creates and stores, and as another example, many medical and scientific institutions create data trusts to store and govern data).
Here’s one way that a Workers Data Trust could operate.
A union (or group of unions) create a Workers Data Trust, where the union’s members would also be beneficiaries and members of the trust. The unions would then negotiate a legal agreement, deed or similar with an employer about the ownership of workers’ data. As part of the agreement, the employer would agree to vest ownership and control of its employees data with the trust. The Workers Data Trust would then set the terms, rules and conditions of the employer’s use of that data.
The data collected could include the workers’ personal and health information, HR information, time and motion data, geo-location data, shift and scheduling data, personal payroll data and biometric data (such as facial recognition).
The Trust’s conditions on data use by an employer could include: access to the data by the worker and their union representatives, restrictions on monetisation or sale of the data, limitations and rules in the use of business analytics, and the physical location of data storage.
Workers Data Trusts could also do what existing medical data trusts do, which is charge usage fees for access to data (the employer would pay the fee). This would enable the Trust to have the resources required to properly and securely manage the data.
When the worker moves to a new employer, the union and Trust would negotiate with the new employer. If there were multiple “industry” Workers Data Trusts, the worker could also shift their data from one Trust to another.
Workers data (as workers, and in other realms of their live) is being commodified and used to control them.
In the near future, the companies that invest in surveillance capitalism will be in an unassailably powerful position to control workers’ lives.
By establishing Workers Data Trusts now, unions can take a proactive step towards shaping data laws that will enable those trusts to have legal powers to represent members and ensure that control over data is held by workers, not corporations.
Just as unions led the creation of superannuation funds to act as long-term (and ultimately powerful) trustees for workers’ retirement savings, Workers Data Trusts will do the same for workers’ data.
Industry superannuation funds in Australia are some of the most powerful financial institutions. The foresight of the union leaders that created industry super funds in the 1980s and 90s led to the establishment of funds that have the potential to influence the corporate behaviour of even massive companies like AMP, Qantas or Wesfarmers.
Workers Data Trusts have the potential to exercise this kind of decisive influence and power — especially as creation, collection, use and analysis of data becomes more valuable.