A View from the CT Foxhole: Air Vice-Marshal Sean Corbett.

• Author: Jahanbani, Nakissa

CTC: Can you tell us a little bit about your background, your military career, and the work you've done over the past several years? Specifically, how did you come into OSINT (open-source intelligence) work?

Corbett: I joined the Royal Air Force in 1988, as a photographic interpreter--shortly before the specialization expanded into the RAF intelligence branch. After my first tour at the U.K.'s strategic imagery analysis headquarters, I went to Belize on an operational tour, which is where I first started working with a U.S. agency, albeit 'informally.' Working alongside and within the U.S. intelligence community has been a theme of my entire career. You tend to find that when you're trusted within the U.S. intelligence organizations and you get enhanced clearances, both sides say, 'Oh yeah, we'll use him in that role because he's trusted and he's a safe pair of hands.' One of my follow-on tours, for example, was as the collection manager in JTF-SWA [Joint Task Force Southwest Asia] in Saudi Arabia--in the U.S. SCIF [Sensitive Compartmented Information Facility]. I've also completed operational tours in Northern Ireland, Somalia, Iraq, Afghanistan, the Balkans, and Libya. I was the U.K. J2 in Afghanistan for a year, working very closely with the U.S. J2 Joint Staff Intelligence, and as the J2 in our Permanent Joint Headquarters [PJHQ] (a)--our unified combatant command, if you like, we only have one. I also have deep experience with signals intelligence (SIGINT), as I ran the military element of GCHQ [the U.K. equivalent of the NSA] where I was responsible for nine locations globally, several imagery analysis-related tours, and some out-of-branch appointments (as you're expected to do to grow career-wise), including as the deputy U.K military representative to NATO at NATO headquarters, and I was also the executive officer for DSACEUR [Deputy Supreme Allied Commander Europe] for a while as well in SHAPE [Supreme Headquarters Allied Powers Europe]. The culmination of my military career was being nominated as the Deputy Director for Commonwealth Integration [DDCI] within the DIA [Defense Intelligence Agency]. I was the first-ever non-U.S. citizen to be a deputy director of an intelligence agency in the U.S. [It was an] incredible experience, of which I am tremendously proud. It was a huge responsibility, extremely demanding, but a huge privilege. Overall, therefore, I've had a pretty varied career; I've loved every minute of it and been lucky enough to have worked with some incredible people.

In terms of how I became an advocate of OSINT, this was more of an evolution over a period of time than a specific trigger moment. For example, in my tours within the J2 at PJHQ, we had to maintain a pretty broad awareness of global security threats, and it was impossible to cover all of them through dedicated, exquisite collection capabilities. But if you asked me for one thing that brought the need to embrace OSINT into stark relief, I would have to say the Ebola pandemic in West Africa in 2014. It was very much an intelligence-led problem set, from understanding its spread and regional impact to the security consequences (reaction of populations, civil unrest, etc.). Almost all of the available information was at the unclassified level, and I think we as a community were not well equipped to operate at this level. To be honest, we really struggled with the problem set. Following the lessons learnt process, we therefore established an open-source cell within the J2 to provide that addition layer of unclassified intelligence.

And the trend towards OSINT has only increased. There's so much happening in the world right now that government intelligence organizations can only focus on so many priorities. You've got to start looking elsewhere for some of your information sources and getting reflections that you wouldn't normally consider. It was really challenging early on, if not for any other reason other than the culture within the intelligence community and a general belief that anything that was unclassified couldn't or shouldn't be used as an intelligence source.

CTC: Can you speak a little bit about what you have done since your retirement in service? Because when we talk about OSINT, there's a lot of interesting components to what you've done during the second part of your career.

Corbett: To be honest, following my DIA appointment, the military in the U.K. didn't really have anything to offer me after that. I'd already smashed through my glass ceiling as an intelligence professional. They say that you know when it's time to move on from the military, and I think that's true. When I returned from the States, I was put in a holding pattern for another two-star role should a suitable one become available, but during that period, I started to take closer look at some of the emerging open-source capabilities. I was lucky enough to be approached by an Earth observation company that was developing some innovative capabilities using video and imagery from commercial satellites, which was really appealing and so I made a fairly rapid transition.

The great thing about that was they gave me the freedom to start developing some technical techniques based on commercially available satellite imagery to support the defense, security and intelligence sector. Specifically, I'm talking about developing machine learning algorithms and artificial intelligence to apply to imagery such as automated object recognition. This can be a real force multiplier for the intelligence analyst. For example, if you are monitoring an airfield with a number of different types of aircraft on it, you can develop the algorithms that could automatically identify not just the numbers of aircraft present, but also the specific types. We were able to do that to a very high degree of accuracy and sophistication, and at the time, this was fairly cutting-edge stuff. Now, that in and of itself isn't necessarily a game changer, but if you can do that at scale with lots of different airfields, then it frees the analyst to focus on what they should be doing and that's focusing on the 'so-what.' For instance, you can set alerts within specified parameters that will flag up anomalies, such as when there is an unusual level of air activity. The analyst can then engage to assess the significance of the anomaly. At that stage, probably the defense community--certainly in the U.K.--wasn't ready to embrace that sort of technology, and certainly not to rely on the commercial sector. But nonetheless, it was good to realize what could be achieved technically and then applied to real-life intelligence problems and I learnt a lot going through the process.

One big lesson was the ability to develop a capability rapidly, something that is extremely challenging in the military research and development community. By sitting down and working closely with the data scientists, coders, and developers so they knew exactly what we were trying to achieve, we were able to meet the requirement rapidly and accurately. Too often in the military community, we fail to articulate the need in the right amount of depth, or remain constantly engaged throughout the process, and are then surprised when the final product falls short.

Another lesson I learnt early on was how many skills are transferrable from the military to the commercial sector. For example, strong leadership, and good communications skills are absolutely essential, as is having a strategy in which a clear vision is translated into an output, where the ends, ways and means are cohered. What are you trying to achieve? How are you going do it? What resources have you got and what additional resources do you need? And then maintaining that aim as much as possible, testing and adjusting when the need arises. This is second nature in the military but often surprisingly absent in the commercial world. In today's information age, another lesson that can be applied both to the miliary and the commercial world is to use AI as a tool at the appropriate moments, but not to be governed by it and make it an end it itself.

I'll always be grateful for my time as an employee in a commercial company, but...