Study identifies attitudes towards personal data processing for national security

Almost 80 per cent of UK adults support national security agencies collecting and processing personal data to investigate terrorism and serious crime, according to new research by the Centre for Emerging Technology and Security (CETaS) at The Alan Turing Institute, in collaboration with Savanta and Hopkins Van Mil, published on Tuesday.

Professor Marion Oswald MBE from Northumbria Law School at Northumbria University, is a Senior Visiting Fellow at the Centre for Emerging Technology and Security (CETaS) and was a co-author on this study. Her research covers the interaction between law and digital technology, and has made distinctive contributions in respect of privacy, fair decision-making and artificial intelligence (AI) and the way that data is acted upon by the public sector, focusing on policing and national security.

A representative survey of over 3,000 UK adults was conducted alongside a citizens’ panel, measuring public attitudes to national security agencies processing personal data. It is the first study of its kind to look at public trust in UK intelligence agencies’ use of data.

The research measured public support for data processing across several different purposes, ranging from investigating individuals suspected of terrorism and serious crime to the use of data to create automated tools for predicting future behaviours.

Public support ranged across these contexts and was not universal. For example, 28 per cent are opposed to personal data being used to train a new automated tool for predicting future behaviours.

75 per cent support national security agencies collecting and processing personal data to detect foreign government spies, and 69 per cent support it to investigate a crime for which they are suspected to be connected.

Other contexts have less public support. For example, only half support data collection and processing to shape long-term strategies and policies of national security agencies, 42 per cent support its use to create automated tools to predict future behaviours, and less than a third support it in the context of sharing with commercial organisations.

Professor Oswald explained: “Our research showed that the public is generally unaware of what powers intelligence agencies have to collect data about them. As a result, many assume that intelligence agencies receive minimal oversight and were therefore surprised and reassured to learn about the hoops agencies need to jump through before collecting data.

“This research comes at a critical time in which emerging technologies are transforming national security data processing. Policymakers and other stakeholders must develop a clear understanding of public priorities and concerns in this area to make informed decisions about whether and how to automate different aspects of national security data processing.”

Alongside looking at the purposes for use of data the research also looked at support for processing a range of datasets for a person of interest to security services or police, including public posts on social media sites, biometric data, identifiable medical data and private text messages among others.

The research showed that support for national security data processing is widespread but not universal, with a sizeable minority opposed across the board. For each dataset tested, more than 20 per cent of UK adults are opposed to national security processing, and more than 25 per cent are opposed to police processing. This applies even for data that is already in the public domain (i.e. public posts on a social media site).

Support also varies significantly across age groups and demographics. Young adults and vulnerable adults are less supportive of national security data processing compared to other groups.

Only half of 18 to 34-year-olds support national security agencies processing public posts on a social media site, compared to 61 per cent amongst ages 55 and over. This is reflected in comparisons of support from vulnerable adults (52%) to non-vulnerable (62%).

The study is also the first to compare public attitudes to human versus machine processing of data in a national security context. Insights from both the survey and the citizens’ panel of 33 members of the public showed that there is no indication that the public finds automated data processing to be inherently more or less intrusive or fair than human processing.

Indeed, a strong appetite for technological innovation in UK national security was identified, with panel members supporting the use of machine-learning techniques provided there are assurances around accuracy, fairness, safeguards and human oversight are involved.

Rosamund Powell, Research Associate at The Alan Turing Institute, said: “This research is the first of its kind examining public attitudes to data processing in a specifically national security context. It demonstrates that there is not just one public perspective on privacy, and that agencies must continue to seek a nuanced understanding of the factors that shape public preferences in order to ensure that data processing remains proportionate in the eyes of the public.”

Sir Brian Leveson, Investigatory Powers Commissioner, said: “In an era of rapid technological change, in which the UK embraces cutting‑edge capabilities to protect the public and strengthen national security, we must ensure these tools operate under the rigorous, legal scrutiny that sustains public trust. I welcome this opportunity to deepen our understanding of public attitudes to UK national security and hope this research will provide a foundation for balanced dialogue and informed decision‑making.”

Professor Oswald leads an interdisciplinary research project, entitled PROBabLE Futures, which investigates the future use of probabilistic AI in law enforcement.

For more information on law at Northumbria, please visit www.northumbria.ac.uk/law