The need to see over the hill and understand the motivation of the enemy remains consistent. AI enabled capabilities will provide a new lens through which to understand the consequence of such issues and predict future action. Operating in a future battlespace will require not only the capability to apply force, however defined, but also the data and data exploitation to understand centres of gravity at all levels of the campaign.
However, the need to understand in order to inform on how we act, must reach beyond the military to connect with civilian supply and logistic chains if we are to protect the critical infrastructure that we depend upon to influence through soft or hard power. In future conflict the front line will no longer be defined as a distant physical location but rather it will be drawn and redrawn at every level at which we support the national objective within a multi-layered networked Internet of Battlespace things (IoBT) system of systems.
For the UK to enable, gain and maintain AI advantage it will need to invest appropriately at scale in AI. AI can be developed across a broad spectrum of air and space capabilities. The approach recommended is to follow continual spiral multiple AI algorithm development to enable, and then gain, advantage. This should then be underwritten with continued investment to ensure that UK’s air and space power is developmentally ahead of adversaries to maintain the advantage.
The UK has, without doubt, established the industrial and academic building blocks to begin to create the structure to produce and deploy world leading AI. However, the country is at a critical juncture in its relationship with AI and failure to grasp this opportunity, potentially because of the challenges it presents, as these blog posts have uncovered may leave the UK lagging behind in the AI arms race
Picking any air and ppace problem set offers opportunities to develop and exploit AI/ML; our key challenge will be to understand where the priority for any development should be. Just because we can, does not mean that we should and that is not just from the trust and explainability perspectives but also from that of capability improvement, recognising that resource will always be limited. There is a need to understand how we define capability performance improvement as well as how we compare the value of different improvements in order to justify which AI enabled systems are developed. Is the potential game changing Air-to-Air combat capability demonstrated by DARPA’s AlphaDogfight more valuable than an AI prognostics capability that reduces maintenance downtime across the fleet of air platforms, and thereby increases mission rates across all mission sets and offers support to the logistics chain?
I argue that a holistic approach to capability development is required, which aims to avoid a cycle of exquisite capability development as opposed to letting functionality drive where the resource is required to meet the challenges of the future. Low cost repurposed off the shelf technologies should sit alongside bespoke capability development but it is clear that the technology should not drive the need in an age that will be defined by a multiplicity of State and non-State actors posing an ever-expanding threat portfolio.
Potential uses of AI in air and space power include:
- Allowing long endurance pseudo-satellites with minimal operator involvement to stay airborne where they are required;
- Enhancing complex air operations across multiple platforms, to maintain control of the air in contested and congested environments;
- Enabling Crewed/Un-crewed Teaming operations, by minimising the additional workload needed in order to manage Un-crewed adjuncts within the battlespace;
- Optimisation of aerial delivery of people, supplies and equipment, ensuring that right supplies are dropped in the right place taking into account the local weather and tactical conditions at the point of delivery;
- Predicting space weather to better inform wider operations;
- Enhancing survivability of platforms, through both proactive and reactive mechanisms;
- Improvements to individual Operator’s training, understanding individual performance and gaps/weaknesses in capability allow for increasingly targeted training;
- Optimisation of tactics, techniques and procedures to improve mission effectiveness;
- Understanding the Human operator and their performance within mission allowing system adaptability and potentially allowing for direct brain-computer interfaces;
- Supporting the cognitive level of humans to step-change increase within Human performance;
- Enhancing multi-domain integration and operations, ensuring the right person has the right information at the right time;
- Improved platform health usage, monitoring and prognostics, to make maintenance more efficient;
- Contributing to the protection of air and space power infrastructure against a variety of threats;
- Command, Control, Communications, Computers, Intelligence, Surveillance and Reconnaissance in both air and space – through processing at the point of collection, supercharging the sensor to effector time scale;
- Collection and analysis and fusion of data from all sources to inform courses of action at speed of relevance;
- Prediction of the outcomes of tactical and strategic approaches to defeat with minimal damage;
- Smart everything! Not just weapons but everything that senses, defends, attacks, supplies, repairs and constructs.
Beyond the more specific aspects of air and space power, there are parallel opportunities to develop AI to support joint action across UK military capabilities as well as wider across the full spectrum of UK capability with significant opportunities to enhance understanding of the prediction of conflicts in the wider world and how they might be de-escalated and or prevented.