Daily catch-up: Defence Intelligence and imagining the 'unlikely but disastrous'

To the Ministry of Defence last night for the ninth meeting of the Strand Group, the public seminar series of the Policy Institute at King's College, London. We were addressed by Air Marshal Phil Osborn, who has been Chief of Defence Intelligence for a year.

Defence Intelligence, which has sometimes been regarded as a backwater tributary of British intelligence, has under his command been upgraded in resources and status. It employs 400 analysts (Osborn was proud that most of those under 45 are women) and with his four-star rank "we are not as vulnerable to head-office iniquities as we used to be," he said.

DI is readying itself to provide real-time monitoring of world threats 24 hours a day, seven days a week. "That means five shifts, which is manpower intensive, but essential if we are not to be surprised," he said. Trying to avoid being surprised is one of the agency's objectives (it is not a separate agency like MI5, MI6 and GCHQ, but a part of the Ministry of Defence).

Osborn had recently conducted an exercise echoing that carried out by Philip Tetlock for the US intelligence service, Intelligence Advanced Research Projects Activity (IARPA), of reviewing DI's analyses over the past 10 years to see how well it had anticipated developing threats. Most of its predictions had been good, but it had underplayed extremism in the form of the rise of Daesh; the extent to which Russia would be militarily assertive; and the problem of migration. This last had been anticipated, but not the "confluence of factors that made it as immediate [a problem] as it is today".

It is not just change but the pace of change that is hard to predict, he said, particularly with autocracies. With slides showing photos of Vladimir Putin and Bashar al-Assad, he spoke of "the mysteries of personality-based grand strategic decision-making".

His next slide was of the "cone of probability", which showed how DI has to try to assess a range from "likely", which is usually "like what is happening now only more so", to the "unlikely but disastrous". DI has to try not just to get its predictions "right" in the Tetlock sense, but to anticipate threats that probably won't materialise but which would be catastrophic if they did.

Naturally, Osborn wasn't specific, but the audience, many of whom were senior civil servants or ex-civil servants, and many of whom seemed to have worked with Osborn and to admire him, enjoyed imagining some of the chilling scenarios conjured up by, for example, the mere title of the "joint cyber and electro-magnetic group".

I came away glad that the government employs so many people to imagine the worst on my behalf.

• Sign up here to receive a daily "Catch-Up Service" email (9am Sunday to Friday) with links to new posts by John Rentoul.