Interview: Anthony Finkelstein, the government's chief scientific adviser for national security

Magazines like The Engineer arrange interviews in two ways. Sometimes we contact organisations to request interviews with people who we think might be interesting, and sometimes they are pitched to us. In this case, the Royal Academy of Engineering got in touch and asked if we’d like to speak to Anthony Finkelstein. He is the chief scientific adviser for national security, we were told. “Basically, he’s like Q from the James Bond films.”

This is not the sort of introduction that one can turn down. Meeting Finkelstein in a basement conference room at the Academy’s headquarters near St James’s Park (a time-honoured rendezvous for Cold War spies, as John Le Carré told us), I wasn’t sure whether to expect the sort of bumbling-but-brilliant character portrayed by Desmond Llewelyn in the original films, the donnish-but-sarcastic type played by his successor, John Cleese, or the unassuming-but-penetrating new version of Ben Whishaw. Finkelstein resembles none of these; he is a tall, slightly gangly man with a taste for floppy caps and warm knitwear, and has a measured manner of speaking that indicates he is considering every word carefully. He is also quite adamant that he isn’t Q. When asked whether there is, in fact, a Q at all, he replied that he can neither confirm nor deny it. “But if there were one,” he added dryly, “she’d be doing a quite marvellous job.”

Finkelstein is a software systems engineer by profession, holding the chair in the discipline at UCL and based at the Alan Turing Institute, the UK’s national institute for data sciences headquartered at the British Library. However, he added, we should not necessarily draw any inferences from that about the nature of the science he advises on. All other government departments save one have chief scientific advisers (CSAs), he explained (the Treasury has a chief economist). “We try to have a kind of mix among all the CSAs of different disciplines, so that when we have a multidisciplinary problem we can call on a range of people. We have physicists and materials scientists and people who have control and robotics experience, and a statistician, so I’m the sort of resident computer scientist.”

Finkelstein works across the range of government organisations that bear upon national security. Formally, his affiliation is to the government office for science, but he deals most often with the Ministry of Defence, the Home Office and the police. His remit covers the whole technology spectrum, but is not concerned with weaponry or what he describes as “a large defence infrastructure”.

The role of CSA is both proactive and reactive. “I run a large national security-related research programme, and I have creative input in that,” Finkelstein said. “But also people in the national security arena bring problems to me which I try to use science to address. The other people who bring me problems are our adversaries; I look very closely at what they are doing and look very closely at how any technologies provide a threat or an opportunity to improve the effectiveness and efficiency of our national security.” The nature of those adversaries is as various as the technologies Finkelstein deals with. “They range from lone actors and spontaneous volatile extremists through to terrorist organisations, or ‘hostile non-state actors’ as they are sometimes called, to states which wish us and our values ill,” he explained.

Although not all of the technologies with which Finkelstein is concerned are in the digital domain, because of his background he has a particular interest in the computer science aspects of national security.

One thing that is particularly interesting to him at the moment, he said, is privacy-enhancing technologies “both for the opportunities they provide for national security to be able to maximise its function while minimising intrusion, but at the same time, I’m really interested in the exploitation of these privacy-enhancing technologies by our adversaries to hide bad stuff”.

Some of these privacy-enhancing technologies include statistical disclosure control, which is a technique intended to ensure that when a survey or administrative data is analysed, no person or organisation can be identified from the results; privacy-preserving data mining, a related technique that tries to ensure that while knowledge can be extracted from data, that knowledge should not include information about individuals or organisations that should be kept private; and another related technology known as homomorphic encryption, which is computing unencrypted data, so inferences can be drawn without decrypting private information.

In the case of such technologies, Finkelstein’s role is to encourage research that will help provide tools for providers of critical national infrastructure to manage related data and ensure that it is kept as secure as is feasible.

"The national security community now knows we can’t do everything behind the barbed-wire fence. We are only going to be able to keep up with exponential technlogy advances and with fast-moving agile adversaries if we exploit the full value of the innovation community and of the open science and technology community"

One thing Finkelstein is particularly keen to stress is that even when security structures seem to be crude, the science that goes into constructing them or optimising them is often considerable. The most visible pieces of the security apparatus that the general public see are probably bollards and barriers protecting sensitive buildings or places where crowds gather. “Designing a really good bollard is extremely difficult because the more deeply you embed it the more expensive it is, and you have to withstand all sorts of force,” Finkelstein insisted. “Actually, vehicle mitigation is really difficult, complex and quite hi-tech. The point is, you always want to use technology appropriately for the challenge you are facing, and that isn’t always the most elaborate solution. It’s the best solution to meet the requirements.”

Another part of security procedures that many people know about is the use of spotters surveying crowds in airports or sometimes in video feeds to identify patterns of suspicious behaviour. “There is an astonishingly complicated bit of behavioural science research that’s gone on in order to underpin changing the way we do things in public spaces. It may appear crude but if it does that belies the fact that it is based on a lot of very serious behavioural sciences.”

As part of ongoing efforts to develop technology in the national security sector, Finkelstein and the Royal Academy of Engineering have launched a post-doc programme to help identify exciting and important research that might make a contribution to the UK’s national security and that of our allies. “We are also interested in developing the UK capability in many of these areas of science and engineering. We want to encourage work on topics which are likely to make us safer, while developing the tech ecosystem around networks to our collective benefit.”

Partly inspired by a similar successful scheme in the US, this programme is a response to a problem which will be familiar to many in the industrial R&D community. “Things have changed, technologies are a lot easier to access and the global science base is a click away, so the government is very conscious that science is globalised, and there is a whole range of states that now have advanced science and are moving at our rate if not faster. The national security community now knows we can’t do everything behind the barbed-wire fence. We are only going to be able to keep up with exponential technlogy advances and with fast-moving agile adversaries if we exploit the full value of the innovation community and of the open science and technology community. It’s difficult and counter-cultural for us to achieve that openness – we prefer in general not to tell people about our capabilities or our lack of capabilities, but on the other hand the brightest people don’t necessarily work for you, so we have to reach out.”