Artificial Intelligence in Warfare. Is your Data Ready?
Last week’s Atlantic Future Forum onboard HMS Queen Elizabeth in New York, brought together senior members of the UK and US military and Government, with innovative companies in the cyber security & artificial intelligence sectors.
Alongside companies such as Lockheed Martin, Raytheon, L3 Technologies, Amazon Web Services, Nvidia, SAP, BlackRock, and the Chertoff Group, Envitia was invited to discuss what role emerging technology will play in defending against threats to the US & UK.
Is Your Data Ready? Moving from Data to Actionable Insight
During the event, it was clear that the changing nature of warfare requires a different approach to that adopted over the past 20 years. The Secretary of State for Defence, Gavin Williamson, told those gathered onboard that:
“I am convinced cyber, information warfare and intelligence must be fully integrated into our military power; that they are fundamental to a modernised defence and our war fighting successes, and that without them we put our nation’s security at perilous risk.”
The key phrase is “fully integrated”. The importance of data to the modern warfighter is fully agreed. Data can come from disparate sources including domestic and allied Government agencies, crowd-sourced databases, social media, and from the huge range of sensors on military assets in the land, sea, air, and space environment, each with billions of data points.
But data on its own is meaningless, overwhelming, and time consuming. And that’s only if data is available. Instead, data is often trapped in “silos” within proprietary systems which prevents integration, collaboration and the potential to transform decision making. This article provides a 5 step approach on how to gather and manage data to create that actionable information required by the warfighter.
5 Steps to become “Data Ready”
- Know what data you have available at your disposal
Every organisation has divisions and departments, with their own systems and applications, that hold data for a specific purpose. That’s fine – it’s normal. However, each application that contains data is just one part of the bigger picture that, if combined with other data types eg. spatial, image, intelligence, operational, business etc., could be useful to other divisions and departments that just don’t know that data exists and how related data could enhance its own decision-making. For example, combining eyewitness reports with signals intercepts and UAV surveillance video would give a richer picture of a live situation and allow better decisions to be made.
- Have a data integration plan
Knowing what data you have, and where, is a great start. Each system or application was designed for a specific purpose – and integrating with other systems would not necessarily have been part of the design. Having an overarching data integration plan, owned by one person, that is shared throughout the organisation with key stakeholder contribution is the next step and essential for planning purposes and coordination.
This will flush out complications around combining databases (or not) and facilitate discussion around the need (or not) for a unified data model & architecture. It will also help with the plan to ensure data accuracy and consistency around open data standards and interfaces so that data is usable across the whole organisation, and by external parties, if needed.
- Find a way to fuse that data
Using the same proprietary system is fine as long as everyone else uses the same system, including every third party and external organisation. Obviously, that’s a great outcome for the owner of that proprietary system, but modern organisations are looking towards interoperability rather than vendor lock-in.
Finding the most effective, efficient, and secure way to fuse or aggregate data from disparate sources will enable data sets to be overlaid with operational and business data. To achieve this, consideration must be given to the use of harmonised data models based on an “open standards” approach (rather than the closed proprietary “black-box”) to ensure interoperability.
However, a word of warning. Because of the nature of data silos within legacy systems, this is the most difficult step of all and often overlooked. The good news is that through the use of emerging data standards, a fragmented data structure is no longer a show-stopper. There is no need to move all data into one database or one physical location, which is expensive, risky from an IT perspective, and exposes the organisation to unnecessary cyber security risks.
- Apply data analytics and artificial intelligence (AI) technology
To stay ahead of adversaries, the modern warfighter and analysts need to find insights and patterns from the huge amount of data at their disposal. The use of data analytics will provide that insight and AI technology, such as machine learning (ML) algorithms, will enable automated decisions where the human brain could simply not assimilate the volume and velocity of data to reach anywhere near the same outcome, and certainly not in near real-time. Take the example provided earlier of a live situation combining eyewitness reports, UAV feeds and signals data – if ML algorithms can learn to detect hazardous or threat situations as they occur, based on certain conditions, then decisions could be made in real-time, ie. to alert troops or first-responders on the ground.
Only once the disparate data sources are available to the warfighter, in an accessible and consistent format, that is accurate and kept up-to-date, will the value of that data increase exponentially and collaboration can start in earnest. Applications can then be developed for a multitude of purposes – each aimed at delivering clear benefit ie. quicker & better decisions or more cost-effective outcomes.
- Re-use data and insights
The result of automated or manual analysis should itself be published and available for re-use to support other decisions. For example, an analyst’s interpretation of a satellite image provides valuable insights that could be re-used with other sources of data to provide further insight, relevant to a separate application. Of course, this works better with a system that uses live data which is updated in real-time so any re-use provides the correct version.
It’s not all about the data. It’s about how to get access to the right data in the right format, at the right time, to collaborate and improve decision making. The above 5 steps may appear simple and obvious but they’re based on the experience of Envitia’s customers. Too often, it’s just too difficult a task within a large and fragmented organisation. Instead, departments do their best to move forward independently to make best use of the data they have available, rather than a whole organisation initiative.
Organisational inertia, legacy systems, and data silos, are the hurdles that are often in the way but should not prevent releasing the value that could be achieved. The current budgetary climate and, as per the Secretary of State quote above, the external threat, means that to look away is not an option. To learn more about how Envitia’s data expertise and data products can help with the above, contact us at firstname.lastname@example.org