The Irony of AI
The Irony of AI
There's a certain irony in organisations paying to learn how to use artificial intelligence while AI itself has emerged from decades of research, scholarship and practical experience developed by countless academics, researchers and practitioners.
That observation isn't intended as criticism. In many respects, it simply reflects how knowledge has always evolved.
Someone observes a problem. Someone else asks a question. Researchers investigate it. Others test the findings, challenge assumptions and build upon the work. Over time, individual ideas become bodies of knowledge, and those bodies of knowledge begin influencing legislation, standards, codes of practice, industry guidance and organisational practice. Organisations translate those expectations into governance frameworks, management systems, procedures and training. Practitioners apply them in the real world, generating new evidence and practical experience that informs the next cycle of research, regulation and improvement.
It's a continuous cycle.
Knowledge has never been static. It has always evolved through curiosity, observation, evidence, application and refinement.
Artificial intelligence hasn't changed that process; it has simply accelerated our ability to navigate it.
The speed with which AI has entered organisations has been remarkable. In a relatively short period of time, we've seen an explosion of courses, certifications, advisory services and consultants helping organisations understand what AI is, how it works and where it can create value. More recently, we've seen Australia's Prime Minister outline a national vision for AI, accompanied by proposals for an Office of AI, national standards and a broader regulatory framework. AI is no longer simply a technology discussion. It has become a conversation about productivity, education, security, creativity, infrastructure, governance, economic growth and Australia's future prosperity.
Much of that work is genuinely valuable. Technology rarely tells us how it should be applied within a particular organisational context. Experience, judgement, ethics and practical implementation still matter, and they always will.
What I find interesting, however, is how quickly the conversation can drift towards the technology itself while overlooking the foundations upon which it has been built.
The large language models attracting so much attention today did not appear overnight. They represent decades of work across computer science, mathematics, linguistics, psychology, philosophy, cognitive science and many other disciplines. Thousands of researchers, academics and practitioners have contributed ideas, theories, experiments and discoveries that have collectively shaped where AI stands today.
The same pattern extends well beyond artificial intelligence.
AI also carries a familiar warning. Good decisions have always depended upon good evidence and, as the old computing principle reminds us, "garbage in, garbage out." AI doesn't change that principle; if anything, it reinforces it. The quality of an AI response is influenced not only by the model itself, but by the information it has learned from, the context provided by the user and the degree of human scrutiny applied to the result. Poor-quality information, weak reasoning or unsupported opinion can be reproduced just as quickly as high-quality evidence. Likewise, poorly framed questions are unlikely to produce meaningful answers.
As more AI-generated content enters the broader information environment, there is also a risk that inaccurate, shallow or poorly supported material is repeated, republished and eventually treated as reliable knowledge. The more noise we add to the system, the harder it may become to distinguish genuine insight from confident repetition.
That places a responsibility on both AI developers and users. We need to consider the quality and provenance of the information we rely upon, provide sufficient context, test important claims against credible sources and resist treating polished language as evidence of accuracy. AI can accelerate access to knowledge, but it can accelerate weak thinking and misinformation just as efficiently.
For safety, risk, assurance and security professionals, that caution is especially important. Poor information can shape risk assessments, investigations, security decisions, control verification and advice provided to leaders. The principle remains familiar: understand the source, test the evidence and maintain human judgement over decisions where the consequences matter.
As safety, risk and assurance professionals, we work within systems that have developed in exactly the same way. The legislation we administer, the standards we audit against, the codes of practice we reference, the investigation methodologies we apply and the assurance frameworks we rely upon did not emerge in isolation. They are the product of decades, often generations of research, inquiry, practical experience, judicial decisions, workplace incidents and continual refinement.
Every serious incident investigated, every inquiry conducted, every piece of research published, every lesson identified and every improvement made has contributed, in some way, to the way we manage work today. Many of those lessons were hard won, often through failure, loss and tragedy. That's worth remembering whenever we rely on the legislation, standards and systems that now guide our work.
Quality management, systems thinking, organisational learning, resilience, human factors, Human and Organisational Performance, Safety-II, psychological safety and integrated risk management have all followed remarkably similar paths. New ideas emerge. Others question them, test them, challenge them, refine them and apply them in different contexts. Some become widely accepted. Others evolve into something quite different. Many eventually influence legislation, standards or accepted industry practice.
That isn't a flaw in the process. It's how knowledge matures. Ideas rarely achieve their full value sitting within academic journals. They need researchers to develop them, educators to explain them, practitioners to test them and organisations willing to apply them in the real world.
The challenge arises when we begin confusing the latest interpretation with the original contribution.
As ideas become increasingly commercialised, it can be easy to forget where they came from. The messenger becomes more recognisable than the researcher. The framework attracts more attention than the evidence that informed it. The summary becomes more popular than the work it seeks to explain. I notice this regularly on social media. Topics such as AI, psychosocial safety, leadership and culture are discussed extensively, often by people genuinely trying to help organisations make sense of emerging ideas. That's not necessarily a bad thing; it's how knowledge spreads. It does, however, remind me of the importance of occasionally tracing those ideas back to the researchers, practitioners and experiences that gave rise to them in the first place.
To me, that makes critical thinking, evidence and professional judgement more important, not less.
AI creates both an opportunity and a responsibility. Its ability to synthesise enormous volumes of information into concise, accessible responses is genuinely remarkable and has the potential to democratise knowledge in ways we are only beginning to appreciate. At the same time, easier access to information can tempt us to consume conclusions without engaging with the thinking, evidence and experience that produced them. Knowledge has become easier to access; whether understanding becomes easier is another question. To me, that's where the real opportunity lies.
The organisations and professionals who gain the greatest value from AI are unlikely to be those who simply automate existing work or reproduce familiar ideas more quickly. They will be those who use it to explore different perspectives, challenge assumptions, connect ideas across disciplines and ultimately contribute something worthwhile of their own.
That strikes me as particularly relevant to our profession.
Whether we're investigating an incident, conducting an audit, reviewing critical controls, analysing organisational risk or advising executive teams, our value has never been in simply finding information. Our value lies in applying judgement, understanding context, weighing evidence and helping organisations make better decisions.
AI can assist every one of those activities. What it cannot replace is the curiosity to ask better questions, the professional judgement to interpret evidence or the responsibility we have to contribute our own thinking back into the profession.
At a time when knowledge has never been more accessible, original thought may become even more valuable. Every idea AI helps us explore today was once an original contribution made by someone prepared to observe, question, investigate and share what they had learned.
The goal, surely, isn't simply to become better at using AI. It's to stand on the shoulders of those who came before us, build thoughtfully upon their work, contribute something meaningful of our own and leave the next generation with an even stronger foundation from which to learn.
