Amazon Replaced a 50-Person Team With 5 Guys and AI in 65 Days, While Bezos Claims AI Won't Kill Jobs

Amazon's chief executive has said an internal rebuild that once required 40 or 50 people for a year was completed by five AI-focused staff in 65 days, sharpening the tension around Jeff Bezos's prediction in Paris last month that artificial intelligence will ultimately create a labour shortage rather than erase jobs.

Amazon had already announced that it would cut approximately 16,000 roles across the company in January. The company said the restructuring was intended to reduce layers, increase ownership and remove bureaucracy, while offering most affected US employees 90 days to seek another internal role.

That makes Bezos's assessment, delivered at the VivaTech conference in Paris on 17 June, difficult to separate from the reality facing workers at the business he founded. 'I completely disagree' with the idea that AI will render human labour obsolete, Bezos said. 'In fact, I believe AI will result in a labor shortage.'

It is a confident forecast. It is also one that lands rather differently when a company is showing how dramatically fewer people can now complete certain kinds of technical work.

Amazon's Five-Person AI Example

Andy Jassy's account of the internal project is the sort of anecdote executives love because it makes AI's promise feel concrete. A job expected to take a sizeable team roughly 12 months was reportedly delivered by five people in just over two months. The productivity gain is not theoretical. It is there in black and white.

But the more awkward question is what happens to the other 35 or 45 workers in that equation.

Amazon did not explicitly state that AI caused its January reductions. In its message to employees, Beth Galetti, the company's senior vice-president for People Experience and Technology, described organisational changes rather than an AI-driven lay-off programme. Amazon said it would continue hiring and investing in strategic areas considered critical to its future.

Amazon is spending heavily on AI infrastructure and automation, while Jassy has previously said that generative AI would reduce the company's corporate workforce over time. A business can call that streamlining, and in boardroom language it probably is. For the people receiving redundancy notices, the distinction may feel like rather thin stuff.

The company's investment is enormous. Amazon's first-quarter figures cited in the source material showed AWS revenue of $37.587 billion (£28.09 billion), up 28% year on year, alongside quarterly capital expenditure of $44.203 billion (£33.03 billion).

Full-year spending was projected at about $200 billion (£149.44 billion). Those figures help explain why markets have broadly welcomed Amazon's AI direction, even as its labour implications remain unsettled.

Bezos's AI Jobs Bet

As technology makes work faster and cheaper, it can unlock demand that did not previously exist. More efficient systems could create new businesses, lower costs and open roles that cannot yet be neatly named.

Bezos argued in Paris that AI would lower the barriers preventing people from building and creating. The result, he suggested, would not be a surplus of idle workers but a shortage of people able to meet the opportunities unleashed by the technology.

Automation has often changed jobs rather than simply deleted them. Yet past industrial shifts are not a clean guide to generative AI, which is moving rapidly into office tasks, software development, design and customer support, areas once thought relatively insulated from automation.

Goldman Sachs estimated that AI was eliminating roughly 16,000 US jobs each month, with younger and entry-level workers facing particular exposure.

That is the crux of it. Bezos may be describing a destination, while workers are stuck in the journey.

Gartner projects AI could begin creating more jobs than it eliminates from 2028 offers some support for the longer-term view. But a worker displaced in 2026 cannot pay rent with a forecast for 2028. Retraining may become essential, particularly for those in routine and task-based roles, yet access to time, money and suitable education is uneven.

Amazon's five-person example is therefore not merely a story about clever engineers and quicker code. It is a preview of how companies may redraw their staffing models before the promised new roles have properly arrived. Bezos may yet be proved right about labour scarcity. The people whose jobs disappear in the meantime will need more than optimism.