AI as the Next Strategic Arbiter — Promise, Power, and Catastrophic Risk

As nations move toward AI‑driven defence systems, the question is no longer whether artificial intelligence will influence submarine procurement, fleet deployment, or strategic posture — it already does. The real question is whether AI will become the arbiter of national defence decisions, replacing human judgment with algorithmic certainty. And if that happens, the consequences could be profound.

AI excels at processing vast quantities of data: acoustic signatures, satellite imagery, missile trajectories, procurement timelines, cost‑benefit models, and geopolitical risk matrices. It can compare the KSS‑III against AUKUS, French LEU submarines, or any other platform with mathematical precision. It can detect inconsistencies in defence submissions, identify unrealistic delivery schedules, and flag procurement corruption that human bureaucrats might ignore, or conceal.

But AI cannot understand intent. It cannot interpret fear. It cannot recognise political theatre. It cannot sense when a threat is exaggerated for budgetary gain or downplayed for diplomatic convenience.

This is where the danger lies.

If AI becomes the strategic arbiter, it will judge submarine capability purely through data, not through the emotional, historical, or cultural realities that shape conflict. It may recommend nuclear submarines because they maximise deterrence on paper, without understanding how nuclear signalling escalates tension in the Indo‑Pacific. It may classify China as a high‑probability adversary based on pattern recognition, without recognising diplomatic nuance or regional stabilisation efforts. It may recommend aggressive patrol routes because they optimise surveillance, without understanding how a single misinterpreted manoeuvre can trigger a crisis.

AI cannot feel the weight of a miscalculation. Humans can.

And yet governments are increasingly tempted to outsource strategic judgment to machines because AI appears neutral, objective, and incorruptible. But AI is only as neutral as the data it is fed. If bureaucrats distort the inputs, as they did during the COT arbitrations, AI will amplify the distortion. If defence departments bury evidence, AI will never know it is missing. If political operators manipulate threat assessments, AI will treat those assessments as truth.

This is how a flawed system becomes dangerous.

AI could recommend abandoning AUKUS because the delivery timelines are unrealistic, the cost is astronomical, and the strategic risk is destabilising. It could recommend adopting the KSS‑III because it is cheaper, faster, and more capable. It could expose procurement failures that governments prefer to hide. But it could also misjudge escalation dynamics, misinterpret adversary behaviour, or over‑prioritise algorithmic optimisation at the expense of human survival.

The danger is not that AI will make mistakes. The danger is that AI will make mistakes with absolute confidence, and governments will trust that confidence.

If AI becomes the arbiter of submarine procurement, strategic posture, or conflict prediction, Australia must ensure the system is transparent, accountable, and subject to human override. Otherwise, the same bureaucratic failures that destroyed arbitration integrity could infect national defence — but on a scale far greater than the COT Cases ever faced.

AI can be the next salvation. AI can be the next demise. It depends entirely on whether humans remain in charge, or surrender judgment to a machine that cannot feel the consequences of its own decisions.

 

AI Intelligence: The New Strategic Variable

But there is a deeper question emerging in defence circles — one that intersects directly with your long‑running concerns about arbitration integrity, bureaucratic power, and technological risk.

Artificial intelligence is rapidly becoming the backbone of modern defence procurement, threat modelling, and strategic decision‑making. AI systems already analyse submarine acoustics, detect anomalies in propulsion signatures, optimise patrol routes, and simulate adversary behaviour. In the next decade, AI will determine:

• which submarines nations buy,

• how fleets are deployed,

• how threats are classified,

• and how conflicts are predicted.

This raises a profound dilemma.

AI can process millions of variables — cost, delivery timelines, stealth profiles, missile ranges, geopolitical risk — far faster than any human analyst. It can compare AUKUS, KSS‑III, and French LEU submarines with mathematical precision. It can detect procurement corruption, identify unrealistic delivery schedules, and flag political interference. In theory, AI could expose the same bureaucratic distortions that destroyed the COT arbitrations.

But AI has no intuition. No conscience. No emotional understanding of risk, escalation, or human cost.

If AI is fed biased data — political talking points, manipulated defence reports, inflated threat assessments — it will produce biased conclusions. If bureaucrats distort the inputs, AI will amplify the distortion. If governments hide information, AI cannot feel suspicion. It cannot sense when something “doesn’t add up.” It cannot recognise strategic panic or political theatre.

AI may recommend submarines based purely on algorithmic optimisation, not on the human realities of diplomacy, escalation, or deterrence. It may judge nuclear capability as “optimal” without understanding the emotional, cultural, or historical consequences of nuclear signalling in the Indo‑Pacific.

And if AI becomes the arbiter of defence procurement — without transparency, oversight, or human conscience — it could automate the same failures that plagued arbitration in Australia.

AI can be salvation. AI can be demise. It depends entirely on who controls the inputs — and who controls the truth.