HMS Warrior’s Ignominy: (AI) Technology Adoption Lessons


Jason Vaughan is an engineer with experience in C4ISR operations and acquisition. He is currently studying for a Masters by Research in Defence Studies with Kings College London, specialising in Artificial Intelligence.  He is also a student on the UK’s Advanced Command and Staff Course.

The cohort of ACSC 22 was privileged to visit HMS Queen Elizabeth during this year’s Maritime Combat Power Visit. As an impressive an engineering and technological feat the Royal Navy’s flag ship offered, whilst on the QE’s flight deck my gaze turned to Portsmouth’s historic dockyard, where I glimpsed the tips of HMS Warrior’s masts. Entering service in 1862 HMS Warrior’s statistics were equally impressive for her time. Displacing 9,210 tons,the Warrior was the Navy’s first iron clad warship, powered by sail and steam her armament made her the perfect deterrent to the French navy. Yet within a mere 7 years HMS Warrior was obsolete and suffered the ignominy of being assigned to coastguard and reserve services duties.

I’m not suggesting that the same fate awaits Queen Elizabeth, or disputing the utility of the carrier programme. Rather, I want to explore the tensions which exist between a culture of technological innovation and maintaining a balanced force structure and operating concept. The annuals of history repeatedly reveal that cutting edge technology is obsolete almost the day it enters service, and underline the necessity of rapid and ongoing adaption in order to keep complex platforms relevant. Yet the same history shows that politicians, policymakers, and military leaders have often been seduced by the ‘newness’ of the latest innovation, and mislead by its apparent potential. How is it possible to maintain a technological cutting edge, without sacrificing important capabilities in order to do so?

An AI Revolution?

The victory of Google’s AlphaGo over the Chinese Go world champion Ke Jie is accepted as common pop trivia, and indeed dominated part of the news cycle for a single day in May 2017.  The victory arguably heralded not just the mastery of machine over man but also recognition of Silicon Valley’s global technology hegemony to rank alongside US economic and military prowess.[I] In military circles the victory is often proclaimed as an example of the impending dominance of AI in the era of Information Age Warfare that western militaries need to react to. Indeed, in many contemporary military AI think pieces, the DOD’s Algorithmic Warfare Cross-Functional Team’s (Project Maven) exploitation of machine vision of surveillance imagery from drones is frequently cited as evidence of an impending AI revolution. Similarly, the DOD’s creation of the Joint Artificial Intelligence Centre is seen as an acknowledgement of the changing character – if not nature – of warfare as a result of the AI revolution.

From Beijing’s perspective, AlphaGo’s victory was regarded as a strategic shock that coined the phrase China’s ‘Sputnik’s moment’ for AI, transforming the cooperation between the Chinese Party State, technologists and entrepreneurs.[ii] It produced a dawning realisation in Beijing that to overhaul US dominance in AI required a commitment of state resources on par with President Kennedy’s vision for the Apollo programme. The construction from the ground up of Xiong’an, the world’s first truly smart (AI), has been the result. The system is expected to be inhabited by 2.5 million people and is attracting state infrastructure expenditure in the order of US$583 billion.[iii]Such commitment is almost incomprehensible to Western audiences but reflects the determination of China to leap-frog the US’s Third Offset Strategy.[iv] Underpinning this endeavour is Chinese recognition that access to data is the primus inter pares for training machine learning, the form of AI that has led to AI’s global explosion that is revolutionising almost all aspects of the Information Age. China’s national AI development plan has set a target for 20% of the world’s data to be created within China by 2020 rising to 30% by 2030. Staggering as such targets are one must consider the defence implications from Chinese endeavours?

Skynet & Military Capability:

To answer such a question we will turn to the Chinese Government’s (unfortunately named) Skynet surveillance system that is benefitting from unparalleled levels of data access.Skynet’s AI is analysing approximately 570 million CCTV cameras nationwide. Far in excess of human comprehension, to run facial recognition software to alert Chinese authorities to the location of individuals as they are tracked across the country. The data is collected by multiple companies and security organisations but the “only owner of the complete, consolidated dataset is the Party-State.”[v] The data collected is then used to improve the performance of the system. Overtime, Chinese authorities believe that Skynet’s AI will be able to predict and anticipate criminal and anti-state activity. Putting aside Orwellian objections arguably the most significant military implication from Skynet’s machine vision is an extensive target recognition capability which is becoming more capable by learning to track a population of 1.39bn people through Chinese CCTV.

If the consequent algorithmic evolutions were transferred onto a military surveillance system, it could radically accelerate the decision-making and target acquisition. Admittedly, legitimate challenges will arise to question how transferable is Skynet’s algorithms to other global surveillance tasks, as the Chinese population are not representative of the entire global population? What is more critical to recognise is AI’s civil-militarily duality for technology development and the Chinese potential to combine multiple Narrow AI systems for military employment.[vi]

Yet even Skynet is likely to rely on some form of assured access to a cloud infrastructure to handle the vast storage and processing of data that AI depends upon to function. From a military perspective the acknowledgement of the contested nature of the electromagnetic spectrum and size, weight and power constraints have all hindered the adoption of AI in the forward battle space, but potentially no longer. The advent of driverless cars and the looming introduction of 5G has focused the research of the Silicon Valley giants that want to exploit the power of the Internet of Things. The traditional approach of communicating with cloud-based AI services is just far too slow where seconds matter for real-time operations. This has led to the development of Edge AI ꟷ effectively the processing of data by AI algorithms on a local hardware devices; the algorithms are using data created on or transmitted to the device.[vii] Edge AI does not need to be connected to a cloud enterprise in order to work; it can process data and take decisions independently without a connection.[viii]The use of smart devices such as Alexa and Siri is driving industry towards reducing power and bandwidth consumption by focusing on the adoption of Edge AI – the environment that the military has to operate in. This is causing strong competition between established and start-up US and Chinese chip manufactures to develop the next range of Edge AI hardware and naturally offers potential for the military to adopt.[ix] Fundamentally, the notion of Edge AI will allow the military or other actors that can enact rapid decision making at the edge of the battle space to dominate an adversary, geo-spatially, spectrally and temporally. Prosecuting action at the edge of the battle space will naturally have a clear tactical and strategic advantage.


So what is the technological lesson to learn from HMS Warriorearly obsolescence as the technology life cycle continues to accelerate? Simply, the necessity to plan strategically for technology adoption is pivotal. Innovation is also crucial but unless successfully adopted into mainstream capability is a wasted opportunity. HMS Warrior’s fate was sealed by a lack of imagination to evolve her capability through incremental upgrades and when necessary transformative innovation – the USAF’s continued utilisation of the B52 is perhaps the clearest example of this approach. As sensors and weapon systems evolve into intelligent machines with embedded AI processing, their ability to prosecute tasks autonomously (human out of the loop) increases markedly. This benefit must be balanced by recognising that Narrow AI (autonomous) systems perform well in low complexity environments with little uncertainty.  Nevertheless, the looming advent of Edge AI (harnessing multiple Narrow AI subsystems) incorporated into autonomous platforms may present an unprecedented opportunity to accelerate the OODA loop to achieve decisive advantage over an adversary.

Image via flickr.

[i]Lee, Kai-Fu., “AI Superpowers China, Silicon Valley and the New World Order.” (Boston: Houghton Mifflin Harcourt, 2018), p.2

[ii]Ibid., p.3.


[iv]Lee, Kai-Fu., “AI Superpowers China, Silicon Valley and the New World Order.” (Boston: Houghton Mifflin Harcourt, 2018), p.133.

[v]Layton, Peter., Royal Australian Air Force Air Power Development Centre, 2018, “Algorithmic warfare: applying artificial intelligence to warfighting”, Air Power Development Centre, Canberra, ACT, p.54.

[vi]Payne, K., “Strategy, Evolution, and War: From Apes to Artificial Intelligence.” Georgetown University Press, 2018, p.167.

[vii]Imagimob., “What is Edge AI?”:, (accessed 15th May 19).


[ix]Venturebeat., “Investors share their predictions for AI and machine learning in 2018”:, (accessed 15thMay 19).


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google photo

You are commenting using your Google account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s