Latest Features
Trending Topics
Trending
This article is part of our Opinions section, where we invite industry professionals to share their views on the most pressing technology questions of our time.
The loudest question about artificial intelligence is also the most personal. What will it do to my job? The usual answers sit at opposite ends of a spectrum. One says AI will replace people wholesale. The other says it is simply another productivity tool. Both are too simplistic.
AI will remove some tasks and some roles. But when technology makes valuable work cheaper and faster, demand often expands around it. The Industrial Revolution did not merely substitute machines for muscle. It multiplied production and created demand for engineers, railways, logistics and entirely new professions. AI is beginning to produce the same kind of second-order effects.
So, the right question is not simply, “Which jobs will AI eliminate?” We should instead ask which work will AI compress, which work it will expand and how organisations should redesign themselves for what comes next?
That tension is especially important in cybersecurity, where AI is often presented as a way to automate overstretched security teams. Yet security operations are not a fixed body of work waiting to be completed more efficiently. It is a continuous contest against adaptive opponents, in which every new capability can expose more risk, generate more decisions and increase the need for skilled human judgement. For security leaders, the real challenge is therefore not simply how many tasks AI can automate, but how to redesign the function around the work that grows as automation improves.
Some tasks are compressible and AI makes them more efficient as it reduces the human effort required or removes the need for a person altogether. Think data entry, routine reporting, basic correlation or administrative analysis. If the volume of work stays fixed, greater efficiency means fewer people are needed to complete it.
Other tasks should be considered as more elastic, as suggested by Geoffrey Hinton, who introduced the concept of AI elasticity. He states that AI actually increases demand for the work around some tasks.
Radiology is a useful example, as initially outlined in Cancerworld. AI was expected to make image analysis so efficient that fewer radiologists would be needed. Instead, faster and cheaper imaging helped increase throughput. As more scans produced more findings, more follow-up investigations, more quality control and more conversations with clinicians and patients was the result.
The work changed, but it did not simply disappear. Expert oversight, standards and accountability remain central to safe adoption. AI can therefore reduce backlogs and improve productivity while increasing the importance of clinical judgement and governance.
The same distinction applies across the workforce of an organization. In sales, AI can compress metrics, research and back-office administration, while making richer customer interaction more valuable. In HR, it can automate benefits descriptions and reporting, while increasing the need for people who can guide employees through disruption. The unit of analysis should therefore be the task, not the job title.
Security operations have effectively unlimited demand as more visibility into growing data streams uncovers more signals that require attention. More signals create more investigations, decisions, remediation and coordination with the business. Even if AI reduces the time required to analyse each alert, it can increase the total volume of meaningful work.
Cybersecurity also differs from fields such as medicine because the problem is intelligent and adaptive. Cancer does not use AI to become better at evading treatment. Attackers do use AI to automate activity, personalise social engineering and alter tactics in response to defences. AI is accelerating both the speed and sophistication of attacks while also strengthening defensive capabilities. That means AI increases the signal and the noise at the same time.
This is why security work becomes hyper-elastic wherever there is an opponent. Routine triage may be compressed, but investigations, threat hunting, engineering, risk decisions and business communication expand. Errors remain expensive, missed signals carry serious consequences and responsibility cannot be delegated to a model. The goal is not to remove analysts from the system. It is to move them towards the work where judgement matters most.
The final implication is organisational as security operations need a revolution comparable to the shift from traditional IT delivery to DevOps. DevOps replaced slow, linear processes with faster iteration, agile teams and continuous improvement. SecOps must now make a similar move.
AI assistants will increasingly handle correlation, enrichment and first-pass analysis. Analysts should no longer be defined primarily as Tier 1, Tier 2 or Tier 3 alert handlers. They will become SecOps engineers who investigate, orchestrate, test, supervise and improve the system around them. Going along with this transformation, career paths will change too. Instead of learning through years of repetitive triage, new practitioners may develop through apprenticeships, mentoring and embedded teams that combine human expertise with AI.
The urgency to adapt comes from the machine speed conflict. Attackers and defenders are beginning to use increasingly autonomous systems. Humans may establish goals, constraints and controls, but the immediate exchange can occur machine to machine. That does not make people irrelevant. It moves their accountability upstream so they must design resilient architectures, validate outputs, set guardrails and decide what a system is permitted to do.
Leaders should begin by mapping work into three categories of compressible, elastic and hyper-elastic activity. They should automate the first, build capacity for the second and treat the third as a strategic capability. Roles, development and accountability must then be redesigned around the reality that AI will generate more output, more decisions and faster consequences.
The future of work will not be defined by a simple contest between humans and machines. It will be defined by whether organisations can compress the right tasks without starving the elastic work that creates value. In cybersecurity, that transition is already under way. The winners will not be those that use AI merely to shrink the SOC. They will be those that use it to build a faster, more capable and more accountable security function.
