AI-driven dismantling of routine tasks breaks the career ladder for beginners in knowledge sectors

Restructuring of large corporations has become an announced executive path driven by artificial-intelligence algorithms, whereas it used to be linked to sudden financial crises. Within a week, the design-software platform Canva announced that more than a fifth of its workforce would be affected by job cuts, role changes or task transfers concurrent with an internal reorganisation around smart models, while Nice, a customer-service software company, laid off about five hundred employees in the same period. The same picture recurs in quarterly reports of several global firms, where executives told investors that the new systems enable cost reduction and operation management with fewer staff, making the visible layoffs merely a surface layer for a deeper shift that infiltrates existing jobs through gradual automation of summarising, sorting and drafting tasks.

Fahd Bazari, managing partner at Bellamy Alden, explains that framing AI replacement in binary terms, either the robot takes your seat or it does not, is inaccurate. The technology does not swallow a job in one go; it takes over partial tasks, which can be identified through a two-question test: first, can the task be taught to a new competent employee? If yes, the models can learn it too. Second, does the task recur continuously? The higher its repetition rate, the stronger the investment incentive to automate it. By this criterion, a high-skill alone is not a shield; a lawyer who reviews opponent documents for contradictions and gaps performs high-level work, yet looks for the same patterns in every case, making the task automatable. This explains research published by Anthropic earlier this year, which showed that the groups most exposed to AI effects are largely those with higher education and higher wages.

Market readings identified five jobs that sit directly on the front line: written work, bookkeeping, translation, first-line customer support, and routine writing. This view aligns with Dr. Sumaya Al-Askari, associate professor of marketing at the Canadian University in Dubai, who includes data entry, administrative support and templated content production in the list, stressing that the risk lies in task automation and changing hiring requirements, not in the disappearance of the professions entirely. Mosata Mati, partnership lead at Women in AI and deputy head of projects at an innovation hub of a major MENA bank, adds that the common denominator of these roles is their reliance on predictable, information-dense tasks such as reading, sorting and verification. ILO estimates indicate that one in four jobs worldwide is now exposed to generative AI models, with the impact concentrated in office work, where 9.6 % of women’s jobs in high-income countries fall in the highest automation risk category versus 3.5 % for men’s jobs.

AI accomplishes the first step efficiently, but it does not guarantee result accuracy on its own.Mati illustrates this gap in the banking sector, where automation of customer-identity verification and transaction screening expands, allowing models to accelerate 70-80 % of execution-task volume, while the remaining share still requires exclusive human judgment, contextual understanding and legal accountability. Dr. Al-Askari cites the UAE business environment as an example, noting that grasping local consumer behaviour and multilingual cultural nuances cannot be reduced to an advertising draft generated by an algorithm. Accordingly, Bazari distinguishes two work patterns: the first is human-led and technology-supported, applying to any activity that demands physical presence, negotiation or personal responsibility for decisions; the second is AI-led under engineering and human supervision, which his firm applied in training programmes by replacing marketing managers with marketing engineers who oversee generative models after their agency cut the time to write proposals by ninety %, from three weeks to three days, while raising win rates.

When we examined hiring pathways, we found that the biggest structural challenge is the erosion of the entry-level rung of the career ladder for new entrants to the labour market. Entry-level jobs consist fundamentally of routine tasks that automation now consumes, depriving graduates of the practice space that hones their expertise through direct work. The signs of this gap have already emerged: PwC in the United Kingdom reduced the number of graduate hires from 1,500 to 1,300 in September 2025 under automation pressure and other factors, and researchers at Stanford’s Digital Economy Lab observed a noticeable decline in hiring of the 22-to-25 age group in the occupations most vulnerable to AI. This creates a future dilemma for employers: if organisations stop hiring beginners because of software, where will the expert talent come from to run organisations tomorrow?

Task redistribution forces talent in the Gulf, Egypt and the Levant to shift rapidly from executor to auditor.In Gulf financial centres that are fast-tracking the integration of intelligent agents into audit, anti-money-laundering and customer-service systems, the competitive balance has moved from speed of data entry or document translation to possessing governance, risk-management and model-output accountability skills. In outsourcing and shared-service hubs in Egypt, Jordan and Lebanon, the falling cost of automated drafting and sorting threatens office roles that relied on sheer volume, compelling universities and HR leaders to redesign training pathways; a candidate’s market value is no longer measured by tool usage alone, but by the ability to financially steer those tools and link their outputs to the complex regulatory and cultural environment of regional markets.

The World Economic Forum’s Future of Jobs 2025 report confirms that 39 % of workers’ current skills will change or lose value by 2030, and although 92 million jobs are projected to disappear while 170 million new roles emerge globally, the difference lies in how productivity gains are invested. The most entrenched domains against automation remain those involving human judgement in crises, field professions, trust-based relationships such as complex sales and counseling, as well as technology governance and cybersecurity. The practical step you must take is to build continuous applied and experimental competence and keep a portfolio that demonstrates your ability to solve complex operational problems in conjunction with AI; the real bet is no longer on protecting an old job but on leading the system that runs it.