At the end of January, Amazon announced it would lay off 16,000 employees. The company described the decision as part of an “anti-bureaucracy” drive rather than directly linking it to artificial intelligence. Still, the timing raised questions. The cuts happened alongside increased investment in AI, and that pattern is becoming harder to ignore.
Across the tech sector, companies are reducing headcount while increasing spending on automation and artificial intelligence. Even when firms avoid explicitly connecting job cuts to AI, the broader shift is clear. Businesses are restructuring around technology that can perform tasks once carried out by people.
This has intensified fears that AI could replace large numbers of jobs and reshape the labour market. One widely cited estimate suggests AI could automate tasks performed by around 11.7 per cent of the US labour force.
It is important to be precise about what this means. AI does not usually replace entire jobs overnight. It replaces tasks. Most roles are made up of several different activities. A marketing assistant, for example, might draft copy, analyse engagement data, schedule posts, respond to messages, and help brainstorm campaign ideas. AI can already handle parts of these responsibilities, such as drafting text or summarising reports, even if it cannot fully replicate the entire role.
However, once enough tasks within a job become automated, companies may decide they need fewer people in that position. That is where job losses begin.
In some ways, this moment resembles previous waves of technological change. Industrialisation and later computerisation both disrupted labour markets before eventually increasing productivity and living standards.
That comparison is not entirely comforting. Earlier technological shifts caused significant job losses in certain sectors, particularly where workers struggled to retrain or relocate. Communities built around single industries often faced the harshest consequences.
What makes AI feel different is the speed. Industrialisation unfolded over decades. AI tools have been adopted across offices in a matter of months. That rapid pace creates concern that workers will not adapt quickly enough. Those most exposed may not be manual labourers, but entry-level white-collar workers, the kinds of jobs many graduates expect to start in.
At the same time, it would be misleading to treat AI as purely harmful. New technologies also create new opportunities. AI could increase productivity, reduce costs, and generate demand in other areas of the economy. In the best case, AI supports workers rather than replaces them, allowing people to focus on judgment, creativity, and interpersonal skills.
Ultimately, the issue is not whether AI will change employment. It already is. The real question is how well society manages the transition. If governments invest in retraining, businesses support skill development, and universities equip students with skills that complement AI, disruption may be manageable.
If that does not happen, AI could widen inequality by concentrating gains among those who own and develop the technology while leaving others behind.
Photo by Nahrizul Kadri on Unsplash