Tech leaders say AI means less work - their staff say they work up to 90 hours a week - BBC

TEXT START: For years now, executives at companies that are pouring hundreds of billions of dollars a year into developing various artificial intelligence tools have insisted that the technology will ultimately mean people will spend less of their time working.

The Dissection

The article exposes a bait-and-switch: executives market AI as leisure while management captures productivity as speed, expanded scope, constant availability, tighter deadlines, and headcount leverage. The reported 70–90-hour weeks are not evidence that AI has failed. They are the transition phase: humans become faster, more heavily monitored, and responsible for checking, integrating, debugging, and improving systems designed to replicate their own work.

The Core Fallacy

The central error is assuming that technical time savings automatically become free time. Under competitive ownership, saved time is reinvested into more output and more aggressive targets until labor becomes redundant. The article also risks treating current overwork as evidence against future job destruction. It is the opposite. Workers are temporary scaffolding around AI systems while firms build tools that reduce their future bargaining power.

The four-day-week promise is therefore not a social outcome. It is an unfulfilled legitimacy story. The surplus exists, but its allocation is controlled by the owners of AI capital.

Hidden Assumptions

• Firms will voluntarily distribute productivity gains as leisure rather than capture them as profit, speed, or market share.
• Increased human workload means AI is creating durable demand for human labor.
• The current implementation phase resembles the eventual steady state.
• Better tools preserve productive participation instead of making fewer workers sufficient.
• The problem is mainly workplace culture, rather than ownership and competitive compulsion.

Social Function

This is a partial truth serving elite self-exoneration and transition management. It correctly documents the hypocrisy and brutality of the current AI race, but frames the crisis mainly as bad culture, excessive sprints, or poor implementation. That converts a structural extraction mechanism into a managerial failure while preserving the comforting premise that AI will eventually benefit workers.

It also functions as an ideological pressure valve: expose the abuse, name the hypocrisy, and leave the ownership structure intact.

The Verdict

The article documents the loading phase of obsolescence, not a contradiction of it. AI initially intensifies human labor because workers must build, supervise, and validate the systems that will replace economically necessary human work. The promised leisure is being swallowed by competition. Once AI achieves durable cost and performance superiority and institutions fail to preserve human-only domains, today’s exhausted builders become tomorrow’s surplus labor.