AI in national security raises proportionality and privacy concerns

A study published to coincide with the Centre for Emerging Technology and Security annual Showcase 2025 event has highlighted some of the concerns the public has with automated data processing for national security.

The UK public attitudes to national security data processing: Assessing human and machine intrusion research reported that the UK public’s awareness of national security agencies’ work is low.

During a panel discussion presenting the research, investigatory powers commissioner Brian Leveson, who chaired the panel, discussed the challenges posed by new technology. “We are facing new and growing challenges,” he said. “Rapid technological developments, especially in AI [artificial intelligence], are transforming our public authorities.”

Leveson noted that these technological developments are changing how information is gathered and processed in the intelligence world. “AI could soon underpin the investigatory cycle,” he said.

However, for Leveson, this shift carries risks. “AI could enable investigations to cover far more individuals than was ever previously possible, which raises concerns about privacy, proportionality and collateral intrusion,” he said.

The CETaS research, which was based on a Savanta poll of 3,554 adults together with a 33-person citizens’ panel commissioned through Hopkins Van Mil, found there is more support than opposition for a national security agency processing data, even for sensitive datasets such as identifiable medical data. The study reported that there is also generally high support for police uses of data, although support is slightly lower for regional police forces than for national security agencies.

However, while the public supports national security agencies’ processing of personal data for operational purposes, people are opposed to national security agencies sharing personal data with political parties or commercial organisations.

Marion Oswald, a co-author of the report and senior visiting fellow at CETaS, noted that data collection without consent will always be intrusive, even if the subsequent analysis is automated and no one sees the data.

She said the study shows the public is hesitant about national security agencies collecting data for predictive tools, with only one in 10 supporting the use of such tools. People also raised concerns over accuracy and fairness.

“Panel members, in particular, had concerns around accuracy and fairness, and wanted to see safeguards,” said Oswald, adding that there are expectations around technology oversight and regulation.

The study also found that despite national security agencies’ efforts to engage more directly with the public in recent years, there is still a significant gap in public understanding. The majority of people polled (61%) report that they understand the work of the agencies “slightly” or “not at all”, with just 7% feeling that they understand the national security agencies’ work “a lot”.

Fellow co-author Rosamund Powell, research associate at CETaS, said: “Previous studies have suggested that the public’s conceptions of national security are really influenced by some James Bond-style fictions.”

However, people are more concerned when made aware of what national security does, such as the collection of facial recognition data. “There’s more support for agencies analysing data in the public sphere like posts on social media compared to private data like messages or medical data,” she added.