Why leaders can learn from spies to build public trust and make better decisions
The former top UK intelligence official reveals how leaders can better make big decisions in a crisis using methodology developed by spies.
- Must organise to be less surprised by surprise itself
- Open debate vital to facilitate bold thinking required
- Safe spaces promote open challenge and ‘truth telling’
- AI and digital technology can deliver the extra insights needed
- Learn lessons from the intelligence world and their innovation skills
This is an edited version of a presentation to the Adarga AI Symposium ‘Language of the Future’ on 22 September. You can watch it at 1’30’30 here.
The COVID-19 pandemic has thrown us into a state of national insecurity.
The pandemic has certainly shaken public trust in leadership’s ability to do the right thing, and to explain their actions sufficiently clearly, transparently and consistently to the public.
Confidence in government to take optimum decisions in a timely manner has been shaken on both sides of the Atlantic. It has not helped that COVID-19 struck on the back of a rising tide of social media misinformation and some Russian information manipulation.
We cannot count on quickly building up trust in the actions of government itself in the midst of the inevitable confusion after a crisis has arrived. Strong, consistent leadership is what will make a big difference here. The need is to generate a sense of purpose in circumstances when danger looms, then to guide the political class and public to reframe their expectations.
Why trust matters
A significant lesson in statecraft to be learnt from current experience is about the value of governments and companies firmly banking in quiet times a reputation for trustworthiness. This should be based on observed, reliable, consistent and truthful behaviour. That means keeping one’s word. We cannot quickly build trust in the actions that affect the public in the midst of the inevitable confusion after a crisis has arrived.
Overall, our nation faces grave problems; some forced on us, some for which we have to accept responsibility. Solving problems requires innovation.
There is much we can all learn from the way that our intelligence community has adopted innovative ways of managing information to keep us safe. The UK is well placed to realise the potential of digital technology and AI across private and public sectors to help us operate effective decision-making systems and earn our way in the world.
As our ADARGA host, Rob Bassett Cross, reminded us, we are being swept along inside an information revolution. It is hard to see where we will end up.
Innovation achieves the unexpected
There are precedents for this. We can all think of science breakthroughs that get translated into technological revolutions that end up turning our society upside down. One of my favourites is Richard Feynman’s theory of quantum electrodynamics. That is the interaction of light and matter.
Who at the time would have thought that would lead to the tuneable laser, and thus to the ability to read and write data at speed, the modern music and entertainment industry and to solid state electronics itself? This enables our satellites, mobile devices, massive processing, the Internet of Things and so much else besides.
Where will we end up next as a result of the powerful combination of data science, rich data and machine learning?
Almost every conceivable kind of information about our environment, the cities we live in, the products and services we use, and about ourselves and our bodies, can - with appropriate error correction - be translated into numerical form. With numbers we can manipulate, search, classify and infer causes from effects using machine learning technologies.
We are in transition. That is certainly true of the UK Ministry of Defence. I see this as a non-executive director of Babcock International. We partner the Royal Navy in the digitisation of logistics and supply. We are starting to use real data from instrumenting usage and wear-and-tear on board ships and submarines, plus running virtual models of equipment onshore to predict when parts need replacing.
Of course, the top of the range car manufacturers have been doing this for years. This feeds the giant design factories in Germany with the information feedstock from real users to improve design and cut costs. Now it is the turn of UK defence to benefit from such technology.
Governments always struggle, but must do better
But why is the public sector often a late adopter?
I understand the frustrations of those like Dominic Cummings [Chief Adviser to UK Prime Minister Boris Johnson] in 10 Downing Street who want to push ahead faster than the system appears to want to. We have been here before.
To underline that point let me share with you a quote from an influential book, Passed to You, Please by J P W Mallalieu published in 1942. It is a polemic against the sluggish ways of the civil service when facing the demands of total war:
“One afternoon in the summer of 1860 a sheet of paper was handed across the table to a Treasury official giving evidence to a Parliamentary committee. The committee had discovered that every one of the 20,000 letters prepared each year by the Treasury was written with pen and ink, and copied in pen and ink twice.
“The committee was shocked. It suggested that as letter presses had been in use in business for a century they might be safely introduced into the Treasury. The Treasury was not impressed. The change would be unfair to the copyists, it would be undignified and anyway it would not work. So the Treasury stuck to its pens and encouraged other government departments to stick to theirs.
“It did not give in to letter presses until 1885; and by that time Remington’s had been selling typewriters for at least 12 years. The same was true for shorthand, introduced into government just as the first Dictaphones were coming onto the market.”
Government now says it embraces new technology. But do we really understand what makes it hard in the public sector, and in some large corporations?
I had a conversation with a senior minister (who had better be nameless). He was complaining at the slow progress in adopting AI technology. I told the Minister that he could find one government department, staffed by civil servants, that delivers world-beating results from digital technology and AI. It is my old department, GCHQ [the UK’s Government Communications Headquarters].
As evidence, I pointed to the disclosures of Edward Snowden. They shocked the public by revealing just how inventive they had been in applying AI algorithms to big datasets to track terrorists and to detect the signatures of malware. I gently suggested to the Minister that there might be three reasons why those civil servants thrived on innovation.
Freedom to think has huge value
The first reason I gave was that GCHQ is based in Cheltenham, a long way away from ministers in London and the interfering long screwdrivers of their special advisers. Of course, the Foreign Secretary - who is responsible for GCHQ - cares deeply about the final intelligence product, and the potential sensitivities of intelligence gathering. But he gives the Department the managerial authority to decide (within the law) where to innovate and experiment.
The second reason is that GCHQ has a privileged relationship with the US National Security Agency. They have always been generous with their support as part of the US/UK shared intelligence enterprise. NSA provides invaluable peer review of the new ideas that bubble up in GCHQ. It is prepared to invest itself in some of them.
The third reason I gave, which is the most important, is having an organisational culture of innovation. The ethos of GCHQ over the past 25 years has been to find ways of doing things with the technology that most people - especially the bad guys - think are impossible. That means all staff understanding what AI and digital technology can deliver.
The traditional British civil service based itself on two principles. First is the humanistic: truth can always be revealed by discussion. The second is the universalist: a person must be capable of understanding anything and of doing anything.
The first principle needs qualification: discussion must be preceded by data, analysis and inference. The second principle still contains a truth: those who have big decisions to take must be capable of drawing on a wide perspective. They must also have sufficient understanding of what the professionals are telling them.
To achieve that, GCHQ recruits the most diverse set of first-class minds it can attract. This includes the traditional double firsts in classics or history, as well as data scientists and mathematicians.
But the difference from most of Whitehall is that there is an absolute expectation that whatever the initial formation of the individual they will put in the hard work after they arrive to learn their digital trade. That is a take-away lesson for all of us: to infuse our colleagues with a comparable determination to keep abreast of what digital technology can offer.
Another take-away from GCHQ is that an ethos of innovation has to be nurtured. In promoting innovation, bullying tactics will be counter-productive. So will letting staff see inconsistency between the words and actions of their leaders.
Keep challenging and asking – without fear
In my experience, problems often arise because policymakers and professionals fail sufficiently to probe each other’s position. The answer a leader gets from the professionals does depend on their knowing the true motivation for asking the question. Just as the good corporate lawyer does not start by telling you what law is, he or she will advise on how to achieve your objective safely within the law.
Prime Minister, Tony Blair once asked me to organise an awayday for Cabinet members and their permanent secretaries. The theme was ‘why can’t we all pull together in peace time the way we do in war?’.
In war you have a common objective, was one answer. Today, we face Covid-19 and Brexit which ought to be common challenge enough. But the confusion we see in government today shows that the spur of crisis is not the complete answer.
Another important part of the solution we identified for Tony Blair was truth telling. This means having private spaces in which professionals, senior officials and senior ministers can get together to speak honestly and directly to each other about what it will take to win. This would not be mediated through special advisers because unwelcome opinions would be left unspoken for fear of attracting displeasure.
That is another take-away: we all need to think about what the equivalent safe space is for our organisations.
We are in a world of strategic uncertainty where we have to organise to be less surprised by surprise itself. Therefore in addition to national security threats, we also need to have effective national security systems to deal with hazards, the action of impersonal forces of nature or major civil calamities.
New principles for progress
All of us need to understand four distinct kinds of information before taking decisions: Situational awareness; Explanation; Estimation; and Strategic notice.
Those four categories form the SEES model which I set out in my new book, How Spies Think.
Situational awareness answers questions about what, when, and where. In a crisis, decision makers at all levels of government and the private sector need to work with the professionals to determine what data, on what common definitions, will be central to their upcoming decisions, and by when it will be needed. Today, that may involve allowing access to sensitive citizen personal data in bulk, and the application of AI algorithms. We cannot take for granted that there will be sufficient public acceptance of digital surveillance and of the use of machine learning in artificial intelligence algorithms, even for the use of mobile phones as Covid-19 alerting instruments. So we need to develop ethical codes for AI applications in which lawmakers, the tech companies and the public have confidence.
Explanation is the second component needed for satisfactory management of a disruptive challenge. It answers objectively questions about “why are we seeing this data?”. For example explaining the vulnerability of some minority communities to Covid-19. As we acquire evidence-based explanations then we can be more confident in moving on to estimating how events may unfold.
Estimates and modelling are central to questions about “what will happen next if we do, or do not, apply some policy intervention?” Estimates will depend on the explanatory model being used, and on the assumptions chosen. There is technology available to make modelling more interactive and intuitive. But we must still expect healthy differences between experts adopting different approaches so there has to be a process that the policymakers will understand for resolving differences. The inevitable trade-offs have to be explicable to the public in terms the citizen will understand. If not, reputations will suffer and conspiracy theories will thrive.
Finally, the policymaker needs strategic notice of possible future challenges to complete the policy process. This helps answer important questions of the “how could we best prepare for whatever might hit us next?” type, or even “how could we pre-empt this risk so that it never significantly materializes?”
When governments fail to act appropriately or in good time, it can be due to professional/policymaker interface problems arising at each of those four stages.
There can be specific “warning failures” that fall into the cracks between adequate analysis and appropriate policy action: looking but not seeing, and hearing but not listening.
There is digital technology available to help crisis managers assemble, weigh up and use what information there is to make probabilistic decisions. But they need to be already familiar with such methods. So I encourage you to start now.
Sir David Omand was the first UK Security and Intelligence Coordinator. He was responsible to the Prime Minister for the professional health of the intelligence community, national counter-terrorism strategy and homeland security. He served for seven years on the Joint Intelligence Committee. He was Permanent Secretary of the Home Office from 1997 to 2000, and before that Director of GCHQ.
Sir David’s new book is How Spies Think: Ten Lessons in Intelligence. It is published by Penguin Viking and is available in hardback, e-book and audio book from 29 October 2020.