Military Technology 05/2021

60 · MT 5/2021 C4ISR Forum There are other challenges, too. The emergent field of open-source intelligence offers methodologies and approaches that promise to trans- form and enhance the nature of the defence intelligence enterprise, but which bring considerable challenges to a resource-constrained, specialist, often rigidly hierarchical and classified field. Adding unstruc- tured, unverified data sets, such as those provided by scanning public social-media feeds, to time-tested, reliable, trusted sources like wiretaps, human intelligence from covert agents, or imagery produced by mili- tary-owned and -operated platforms, will bring huge benefits in aware- ness and understanding. But doing so raises complicated questions for institutions to address, both from the verification and validation perspec- tives (which sources can be trusted? Under what circumstances? And to what degree, compared to traditional data channels?) and from workload, resource-management and personnel angles. Demand for better intelli- gence, delivered ever more quickly, is pointing militaries in one clear di- rection: toward automation. “Open source offers huge advantages when it comes to intelligence sharing and protecting our most precious covert techniques,” Hockenhull said. “But if you talk to our analysts, many will wince at the idea of even more information and intelligence. What they tell us they need is a way to manage the deluge of data. That is why we’re investing heavily in auto­ mation […] I use the word automation advisedly, because there some- times is too much hype around what many are selling as true artificial intelligence […] Much of what we need at the moment, in defence in- telligence, requires more mature tech capabilities such as data science, machine learning, and automation. As real artificial intelligence matures, I know it will play a vital role in our future, but it will also bring with it a host of challenges around exploitability, fragility and ethics. Our focus, at least initially, within defence intelligence, will be on exploiting the more mature technologies, while ensuring that we play our part in helping shape future ones.” These points were echoed in another keynote at the event, from Juliane J Gallina, Chief Information Officer of the US’s foreign-intelligence service, the Central Intelligence Agency. She recalled the famous landmark in the development of so-called artificial intelligence, where the IBM computer Deep Blue beat chess grandmaster Gary Kasparov; but also pointed out that, later, Kasparov worked on pairing human chess grandmasters with automated systems, and those combined teams beat computers every time. It is this area - the teaming of human analysts with ‘intelligent’ com­ puter systems – which the Agency is looking at to shape future capa­ bilities. These ‘centaur warfare’ capabilities promise transformational benefits, moving intelligence agencies beyond analysing past events and into the realm of, perhaps, being able to predict future ones. “Since the idea of artificial intelligence has become really popular, people will often say to me, ‘Gosh, it’d be really great if we could have artificial intelligence finally get us to being predictive’,” she said. “And I want to say, don’t stop there. Artificial intelligence is not synonymous with prediction […] You can walk the floor of the demo centre here, and you can see many, many industry partners who have really done a fabulous job in common operating pictures, combining the dots on maps, under­ standing the terrain, and making sense a little bit of what’s happening […] But that’s not enough. Because in order to really maintain your advantage, you really need diagnostic analysis. You need to know why did this hap- pen? You need to know, maybe, what could happen next? That’s where the predictive or anticipatory analytics come in. And then, ultimately, op- timisation in analytics is - how can you influence that? What do you do to marshal in your resources? What’s the course of action?” “My challenge to industry [is to] help us embed artificial intelligence in every part of the cyber terrain […] We need human-machine teaming throughout the spectrum, ultimately to help us with course-of-action recommendations, and I’m really challenging all of our industry partners to think about how to inject machine learning into every part of that analytic food chain.” financial data before you try and hack in to Dstl [to see where money is being spent]. I’m not telling anybody any secrets there – that’s what we do to each other. We’ve got to look at data in the round, and look at how we protect it, how we make it resilient, and also deceive using our own data so that people draw the wrong conclusions.” The Royal Navy, too, is working out answers to the questions posed by the need to communicate better, understand more, and do it all securely. Technologies suitable and empowering for society as a whole may not always be viable for military use, even if the speed of manoeuvre and low cost of access make them tempting. “Cloud [computing] is a fabulous capability: I’ve got 280,000 users, four million devices on my networks, across four classifications, and cloud works for the vast majority of that Venn diagram,” stated Rear Adm Nick Washer, Director of Operations at Defence Digital, part of the UK’s Strategic Command, during a panel discussion on cyber’s impact on in- tegrated operations. “It’s not so good necessarily for those people who are at range, disadvantaged users, or 60 metres dived. So, how do we support them from a cyber perspective too? How do we defend them from a cyber perspective, and how do we assure their communications as well? We need to make sure that we are as cyber defended in our satellite reachback as we are in our HF infrastructure.” The next generation of UK military satcom will arrive in the middle of this decade with the launch of the first SKYNET 6 satellite. The capability will be managed by the RAF’s newly-formed Space Command. (Photo: Airbus)