41% of companies expect workforce reductions tied to artificial intelligence in the next five years. The World Economic Forum found that figure, framing a broad corporate expectation of job cuts as firms chase efficiency gains. Companies are already automating repetitive tasks and pruning entry-level roles, while evidence from GitLab and reporting in Forbes shows demand shifting toward people who combine technical fluency with judgement, context and leadership. The contrast matters: firms may shrink numbers, but not all jobs are the same kind of replaceable.

Households, businesses and the technology sector stand to be the clearest near-term victims of that shift.

Where automation bites

Companies have begun to target the most predictable parts of work. Salesforce instituted a hiring freeze for software engineers in 2025, a signal that some employers are slowing headcount growth even as they continue to invest in artificial intelligence. At the same time, the GitLab Global DevSecOps report found that less than a quarter of developers' time is spent writing new code, with the majority devoted to maintenance, testing and security tasks. Those are precisely the tasks where AI can introduce efficiencies.

Multiple analyses agree on the pattern. AI excels at prediction, pattern matching and routine execution, so it substitutes most readily for repetitive, well defined tasks. Forbes documented warnings from AI industry leaders that entry level and routine white collar roles will be hit first. The consequence isn't just a reduction in certain jobs. It's a hollowing out of the traditional pipelines that once trained future managers.

That pipeline effect is the risk many commentators stress. When entry level positions disappear or are reshaped into shorter, more automated stints, organisations no longer get the extended hands on experience that builds judgement and institutional memory. Forbes and trade reporting point to cases where removing those roles has left firms with weaker oversight and less credible leadership succession.

Why learning still pays

That distinction between routine work and higher order activity explains why people who keep learning remain difficult to replace. Commentators with computer science backgrounds emphasise that generating code isn't equivalent to the broader disciplines of computing.

Mastheads

Systems design, engineering complex infrastructure, cybersecurity, verification and domain judgement are activities where stakes and unknowns matter and where current AI can't reliably replicate human judgement.

Human resources and leadership literature adds another layer. Skills such as empathy, contextual judgement and institutional memory are hard for models to mimic. TaQuonda Hill, a senior information technology transformation leader quoted in Forbes, put it plainly: "If you strip out the context, you gut the leadership pipeline. You might gain speed, but you lose the lived experience that turns managers into leaders." The point isn't sentimental. It's about safety, oversight and the capacity to manage novel risks that emerge when systems scale.

Industry studies set out two linked labour market dynamics. First, automation tends to reassign human time away from low skill repetitive work toward higher value activity.

GitLab's findings suggest more than three quarters of developers' daily work is susceptible to automation or augmentation, creating space for engineers to focus on design, security and innovation. Second, employers will increasingly prize blended skill sets: technical literacy with domain expertise, the ability to orchestrate AI tools, and interpersonal capabilities that support team development.

The practical implication is straightforward. Organisations that automate without preserving experiential training may secure short term savings but weaken long term leadership and risk management capacity. Analysts argue firms should rethink how to staff and train for leadership if they intend to replace routine tasks with automation. The talent pipelines that once produced middle managers and senior engineers were not incidental costs. They were investments in institutional competence.

For workers, the message is similarly plain. Continuous learning and on the job experience accumulate a form of value that current AI can't substitute. Technical fluency remains essential, but it must be combined with judgement and contextual knowledge that comes from sustained exposure to complex systems. In short, the safest labour-market position isn't a single skill. It's a pattern of updating skills while deepening domain experience.

That reading is supported by the most concrete recent signals. The World Economic Forum's 41 percent statistic frames employer expectation for reductions.

GitLab's DevSecOps report provides detail on the composition of developer work. And corporate behaviour such as Salesforce's 2025 hiring adjustment shows employers are already changing patterns of recruitment. Together they point to a labour market that will shrink some job counts while heightening the value of learners and leaders.

Related Articles

Keep the World Economic Forum's 41 percent figure in mind: firms are preparing to cut roles, but the work that remains will reward continuous learning, judgement and systems expertise.

This article was created with AI assistance.