The UK´s Open Data Institute (ODI) shares its experiences working to develop a new model for data access, based on the concept of a trust.

Over the years, different data access models have been proposed in order to address cases when data can’t be easily published due to legal, operational or social constraints. There are many ways – such as data cooperatives owned by their members, or data research labs –  to facilitate access to data that might otherwise stay closed, so that its value can be harnessed by specific groups and individuals. But how can this access to data be scaled in a way that maximizes benefits yet retains trust?

The ODI is looking into this question and the various models that exist. One of the newer forms of data access models we have been looking at is data trusts.

What is a data trust?

We first explored data trusts in our innovation programme, funded by Innovate UK. We initially found multiple broad definitions of data trust, ranging from legal structures to storage mechanisms to public oversight schemes. We did not think maintaining a range of definitions was useful – it could mislead people and risk wasting effort due to talking at cross purposes. As a result, we have adopted the definition of a data trust: A data trust is a legal structure that provides independent stewardship of data.

We are now working with the UK Office for Artificial Intelligence, Innovate UK and other partners in the UK and abroad to explore whether data trusts are effective for increasing specialized access to data for the purpose of addressing specific industry, societal and environmental challenges around the world. We are currently running three very different data trust pilots which aim to tackle three diverse issues: the illegal wildlife trade, food waste and city services. The aim is to help with those issues, develop a blueprint for a data trust and make recommendations for what they could become.

How do data trusts work?

Historically, trusts have been used in law to hold and make decisions about assets such as property or investments. A data trust takes this concept of holding something and making decisions about its use, but applies it to data. It is a legal structure that provides independent stewardship of some data for the benefit of a group of organisations or people. That benefit might be to create new businesses, help research a medical disease, or empower a community of workers, consumers or citizens.

In a data trust, the individual or organisations that own or hold data take the role of “trustors”. They grant some of the rights they have to control the data to a set of trustees, who then make decisions about the data – such as who has access to it and for what purposes.

The beneficiaries of the data trust include those who are provided with access to the data (such as researchers and developers) and the people who benefit from what they create from the data.

The trustees take on a legally binding duty to make decisions about the data in the best interests of the beneficiaries. This is sometimes referred to as a fiduciary duty. Proponents of data trusts suggest this duty would help to increase the trust that individuals and organisations have in the way data is used.

Cases for using Data Trusts

Our three data trust pilots, which will conclude in a few months later this year, cover very different sectors and issues, but the core questions we’re looking into answer are the same: user needs, legal structures, technical approaches to name a few. We are working closely with legal firms and community engagement specialists to ensure we have the right expertise to give us good insights both for these pilots and for the future of data trusts.

Tackling the illegal wildlife trade around the world

One of the pilots aims to help reduce illegal wildlife trade by making wildlife data from across the world more accessible. This pilot is exploring data from those working in the wildlife community, specifically data which has the potential to help end the illegal wildlife trade. Working with experts in conservationism and tech from WILDLABS, the pilot will focus initially on two areas where the sharing of data within a data trust could be used to help with machine learning and AI.

One element of the pilot will look at whether image data of endangered species can be made more accessible and used to train recognition algorithms for border officials, helping them to identify illegal animals and animal products.

Secondly, the project will look at image and acoustic data that could identify animals or people moving through protected areas to see whether we can increase access while protecting privacy. Algorithms within remotely placed acoustic sensors could be trained to detect gunshots fired in protected areas, or illegal fishing vessels coming into protected waters, for example.

Reducing global food waste

The Food and Agriculture Organization estimates that food loss and waste results in a global economic loss of $940bn each year. Understanding how much food is wasted helps policymakers and regulators make informed decisions, and reducing food waste saves manufacturers and retailers money, while lowering greenhouse gas emissions. This pilot will look at how food data – particularly data on the nature of food waste and where it ends up – can help track and measure food waste in supply chains.

Improving city services

In the last pilot, the ODI is working with The Royal Borough of Greenwich and the Greater London Authority to look at whether data collected through the Sharing Cities Programme could be made available in a data trust to bring about benefits for citizens. The data could be about energy consumption, collected by sensors and devices in buildings, or about parking spaces and charging bays for electric vehicles, which would inform products and services that citizens can use as they navigate their cities.

Potential benefits and drawbacks of a data trust

One motive behind establishing data trusts is to distribute the benefits arising from data more equitably. In some cases the benefits could be monetary – for example a share in the profits generated by services created from the data. Some people want to create data trusts to create collective power over data, for example by helping a group of workers have more control on data about their jobs, or a community have more control of data about the place where they live. Other benefits would be indirect and difficult (or impossible) to distribute back to trustors, such as the societal benefit of helping researchers understand how to manage mental health issues.

It’s also necessary to consider how data trusts could be used in harmful ways. In the same way trusts have been used to avoid taxes, there is concern that data trusts could be adapted to try to obfuscate the profits generated by data or avoid data protection responsibilities. They will need to be transparent and work openly so that regulators, the public and others can hold them to account. We have become much more aware of who has access to data – data about ourselves, our family, our friends and our work. While we see many benefits from the use of data, such as being able to plan a train journey quickly and easily with an app using route and timetable data, there has also been misuse and harm, as we saw in the case of Facebook and Cambridge Analytica.

 

In conclusions, there are many potential models for providing access to data. Each of them are useful in different and occasionally overlapping contexts. Data trusts are a potential new way to help realise the benefits while preventing the harm. We’re keen to explore them to find out where they might be useful both in the UK and around the world.

 

If you’d like to find out more, get in touch at info@theodi.org

 

By Peter Wells, Jack Hardinges, Lawrence Kay from the Open Data Institute