The Cambridge Analytica scandal has sparked a backlash against the use of personal data, but research suggests the public will still support data science projects that they believe will benefit society.
Dr Mhairi Aitken made the discovery during the course of her research into how people feel about the use of their data.
Aitken is a fellow for the public engagement team at the Farr Institute, a UK-wide research collaboration involving 21 academic institutions and health partners, including her employer the University of Edinburgh. The team has been investigating public opinion through facilitated group discussions between a range of participants.
"We found that confidence that a project is likely to bring about public benefit is absolutely crucial for people to support it," Aitken told Techworld at DataFest in Edinburgh.
"If they don't consider it to be worthwhile, they're not going to support it, whereas if they do think it's worthwhile, they're more inclined to be less concerned about the other aspects."
In 2016, Aitken and her colleagues published a systematic review of the international literature on public views on healthcare data.
The study revealed widespread support for using data in research, but only under certain conditions.
People wanted assurance that their confidentiality would be protected, that safeguards would be in place to prevent the misuse of data, that data would be handled securely and that they would have a sense of control over how it's used.
Most importantly of all, they needed to be confident that their data will bring public benefits.
They would likely support any project that would benefit society even if they weren't sure that all the other conditions would be met.
But if they didn't believe that the project would yield public benefits, they would be highly unlikely to support the initiative even if they felt all the other conditions would be met.
Aitken's research refutes previous studies that claimed public views on data are primarily self-interested.
"The concerns aren't so much about risks to the individual," she says. "They're more about what this data science is for and what it's all about. Are people just doing it for profit and to make commercial companies very rich, or are they doing it to add benefit to society?"
The use of public sector data causes particular concerns as the results could influence policies around health, social care, housing and welfare that reflect the characteristics of groups rather than individuals.
They also worry that this could lead to discrimination against certain sections of society.
Underpinning all of their views is trust. The extent to which people trust the organisation using their data will determine whether they believe the claims about a project's benefits.
"Thinking about what the public benefits of a project would be and how we communicate those public benefits is really crucial for maintaining public support," says Aitken.
What is the public benefit?
The public benefit is easier to understand than define. People often explain the benefits to Aitken through examples such as the Tesco Clubcard, which takes their data in exchange for the direct and tangible result of shopping vouchers.
Larger data science projects can have more opaque outputs, which makes it harder to elucidate their benefits.
When Aitken asks people who work in data science about the public benefits of their projects, they frequently talk about the economic impacts of creating new industries and creating new jobs in data-driven industries.
When she discusses the subject with members of the public she rarely hears these arguments.
Instead, they talk about the more direct impacts, such as using public sector data to improve public services, or healthcare research to develop new medical treatments.
It's also hard to define which members of the public should benefit. Should they be from a neighbourhood or a nation? Should you help the maximum number or the most disadvantaged? Should it aim for a big impact on a small number of people, or a small impact on a large number of people?
These questions have different answers depending on the project and the people you ask, which is why it's important to seek the views of diverse groups of the public.
"If you bring together people from different backgrounds they do tend to think about it in that more socially minded way," says Aitken.
When Aitken began her research eight years ago, it was hard to gain interest from the public in a subject that most people found largely difficult to understand and irrelevant to their lives. That changed as data scandals began to draw headlines, from the Care.data patient information-sharing scheme to Cambridge Analytica's harvesting of Facebook user data most recently.
Public engagement doesn't only help avoid such controversies. It also offers an opportunity to improve projects by understanding different views and interests.
"It's not about making the public trust us, it's making sure that we are trustworthy," says Aitken.
This means listening to people express their concerns and preferences and then reflecting them in data science practices.
Aitken believes that data science tends to focus on what can be done with people's data, and instead needs to focus on what should be done.
"Go into it enthusiastically and openly and don't try and anticipate what people are going to say," she suggests. "Go with questions rather than answers, and be open to being surprised about what people might tell you."