I was chatting recently with one of Accenture’s deep experts in Artificial Intelligence (AI), and he made a comment that I felt was particularly relevant to defence. He said that at its current level of maturity, AI is capable of outperforming humans in execution-level activities–but not in strategic thinking.
This view echoes something that I’ve thought for some time: AI isn’t “streetwise”. By this I mean it lacks sufficient experience of how the world works to think its way out of difficult or unprecedented situations. In this sense, AI is actually quite naïve–a bit like a country cousin who’s new to the city.
Why is this relevant to defence? Well, at one level it puts into perspective the dire warnings in the popular press about AI-enabled “killer robots”. AI machines are just that: machines. They’ll learn what people teach them. And there’s no prospect of them overtaking us any time soon in our ability to think and respond to new and unexpected scenarios.
But the fact that AI isn’t street-savvy also brings some deeper implications for defence. Because it helps to guide us to those areas of the military ecosystem where AI will have biggest and most immediate impacts in the next few years.
To explain what those areas are, let me start by looking at the wider landscape of innovation in defence. Four or five decades ago, defence agencies led the world in technical innovation: the relentless pressure to innovate–combined with plentiful funding from governments–enabled them to break new ground continually, often leaving the private sector trailing.
Today the polarities are reversed. Strains on government coffers, and a rapid migration of capital and talent to the fast-growing internet players, have seen the private sector streak ahead in the innovation race. And AI is one of the areas where the private sector’s lead is biggest.
The effect is that even the investments by the handful of heavyweight military spenders among the world’s governments pale in comparison to those being made by the private sector. For these countries and for smaller nations with more limited defence budgets, the best option is to create a defence ecosystem in which they can draw cost-effectively on AI innovation and talent from the private sector.
But will they do this? My interactions with defence chiefs across the world confirm that many are still skeptical about AI’s application. However, my view is that investing in AI is not just an option–but an absolute necessity.
Why? Because there are a number of key areas where a failure to invest in AI will leave any defence force seriously behind the curve in the next few years. And taking into account that AI’s sweet spot is essentially executional rather than strategic, three areas spring immediately to mind.
The first is cyber security or cyber defence. The processes where AI is most relevant to any organisation are those that are time-sensitive or require hyper-precision–and cyber is both. As recent events have underlined, the scale and speed of today’s cyber attacks put them beyond the capabilities of human beings to comprehend, let alone respond.
It follows that any country facing an AI-enabled adversary–and evidence shows that many already are–cannot afford to rely on human resources alone. It must have AI cyber capabilities itself, or it’ll inevitably be outflanked. So, using AI in cyber defence is a must.
The second area is defence logistics. For any military force, maintaining a real-time dashboard that shows–say–the location of all its tanks and the related munitions and parts is a very complex task. Add in managing the end-to-end defence supply chain to the front line, and the need for AI becomes even clearer. What’s more, shifting the burden of logistics onto AI frees up trained military staff to do the job they’re actually qualified for.
The third area for AI is international collaboration environments. Today’s multilateral missions can include forces from up to 20 or more nations. Sharing information seamlessly yet securely between these partners to create joint situational awareness is a huge challenge. AI is a natural fit for this role: for example, it can help to monitor network traffic and determine network priorities and security classifications, as well as fusing together multitudes of data and visualising it all in an effective way.
In this blog, I’ve looked at how AI can enable defence forces to “do things differently”. However, it also opens up opportunities for them to “do different things”, by undertaking activities that simply weren’t possible before. In my next blog, I’ll look at some of these opportunities. Stay tuned.
See this post on LinkedIn: AI in Defence. Rank: private.