When advice carries enough weight that people stop questioning it, who pays the price? Christian Hill, Group Managing Director, Intelligent Protection International Ltd. shares his perspective.
AI is already influencing decisions across the security sector. It analyses intelligence, identifies threats, assesses risks, recommends responses and increasingly provides guidance on what actions organisations should take.
From geopolitical forecasting and cybersecurity to business continuity and crisis management, AI is becoming a trusted adviser to governments, military planners, security professionals and corporate leaders alike.
But perhaps the most important question is not whether AI is capable of providing advice. The more important question may be what happens when that advice begins to carry so much weight that people stop challenging it…
The new adviser in the room
Throughout history, decision-makers have relied upon experts, advisers and intelligence assessments to help navigate uncertainty.
The arrival of AI has introduced a new type of adviser: one capable of processing vast quantities of information, identifying patterns invisible to the human eye and producing recommendations at a speed no analyst could ever match. The advantages are obvious – but the risks may be less so.
Most discussions surrounding AI focus on what the technology can do. Can it predict conflict? Can it identify threats? Can it recommend the most effective course of action? Yet the real issue may not be the capability of the technology itself. The real issue may be how humans respond to it.
Will AI remain a tool for guidance and insight or will decision-makers gradually begin treating its recommendations as authoritative? In many respects, this process may already be underway.
Recent conflicts have demonstrated how rapidly AI-assisted intelligence analysis is evolving.
Intelligence agencies, military planners and security organisations increasingly use AI to process satellite imagery, open-source intelligence, social media activity, cyber-indicators and economic data at a scale that would be impossible for human analysts alone.
The result is faster threat identification, improved situational awareness and increasingly sophisticated risk forecasting. This means the other question is no longer whether AI can contribute to security decision-making – it is how much influence its recommendations should ultimately have over the decisions that follow.
Today, organisations routinely rely on technology to support decision-making.
Risk management platforms assign threat scores. Security systems prioritise incidents. Predictive analytics identify emerging trends. Cybersecurity tools recommend defensive actions. AI-driven intelligence platforms increasingly provide assessments regarding geopolitical instability, civil unrest and potential conflict.
In theory, these systems are designed to support human judgement. In practice, there is a natural tendency to place trust in recommendations that appear objective, data-driven and impartial.
After all, if a system has analysed millions of data points, considered hundreds of variables and produced a recommendation based upon evidence that no individual could realistically review themselves, who are we to disagree?
Imagine a multinational organisation operating in a politically unstable region.
An AI platform identifies an increasing probability of conflict and recommends relocating personnel, increasing security measures and activating contingency plans. The recommendation is supported by satellite imagery, military movements, economic indicators, social media analysis and intelligence reporting.
Would decision-makers follow the recommendation? Many probably would.
Now consider the alternative. The recommendation is ignored and six months later the conflict materialises exactly as predicted. The questions that follow are entirely predictable.
Why was the recommendation ignored? Who made that decision? What justification existed for disregarding the warning? Faced with such scrutiny, many leaders may conclude that following the recommendation is the safer option, even if they are not entirely convinced by it.
This is where things change. The conversation moves from, “What does the AI recommend, and do we agree?”, to, “The AI recommends this course of action. Is there any reason not to follow it?” The difference may appear small, but its implications are significant.
When people begin searching for reasons to support a recommendation rather than independently evaluating it, the role of the decision-maker also starts to change. The danger is not necessarily that AI becomes more intelligent than humans – the danger is that humans become increasingly reluctant to challenge it.
Statistically sound – yet strategically flawed?
This concern extends far beyond the security sector. History is filled with examples of individuals placing excessive trust in systems that appeared authoritative. Financial models have failed.
Intelligence assessments have proved incorrect. Risk calculations have underestimated threats. Expert consensus has occasionally been wrong.
The common factor is not the technology, but the human tendency to assume that confidence and complexity equate to accuracy. AI is not immune from these limitations. AI systems remain dependent upon the quality of the data they receive. They can be influenced by incomplete information, inaccurate assumptions, flawed methodologies and deliberate deception.
They are capable of identifying patterns and probabilities, but they do not understand political motivations, personal ambitions, cultural influences or the countless human factors that often determine real-world outcomes.
A recommendation generated by AI may be statistically sound and yet strategically flawed. It may be operationally efficient but politically unworkable – it may be logical on paper but entirely inappropriate in practice. This is particularly relevant when considering national security, military operations and international crises.
When advice becomes authority
As AI capabilities continue to advance, there will undoubtedly be pressure to place greater trust in its recommendations. The promise is compelling: faster decisions, improved accuracy, enhanced situational awareness and reduced human bias. However, there is an important distinction between informing a decision and making one.
The most effective organisations of the future are unlikely to be those that blindly trust AI, nor those that reject it outright. They will be the organisations that understand its strengths, recognise its limitations and maintain the confidence to question its conclusions when necessary.
AI may become one of the most valuable advisers ever created. However, it remains an adviser. The responsibility for decision-making must rest with those accountable for the consequences.
The greatest risk may not be that AI becomes intelligent enough to make decisions for us. The greater risk may be that we become comfortable enough to stop questioning its advice – and that may be the day when advice becomes authority.
About the Author
Over a career spanning more than three decades, Christian has worked across military service, international security operations, crisis management and strategic risk, supporting organisations, royalty and high-net-worth and ultra-high-net-worth individuals in diverse and challenging environments around the world.
Intelligent Protection International Ltd.
