AI productivity gains are becoming a new way to demand more work
The AI productivity trap is the pattern in which tools that speed up routine tasks are used not to reduce working time or stress but to justify higher workplace expectations, expanding workloads and intensifying burnout rather than creating genuine time savings for employees. Artificial intelligence now promises to handle tedious work at unprecedented speed, yet in many offices any minute saved is quickly swallowed by workload creep and pressure to produce more. Cybersecurity researcher Keith Jones illustrates the upside of AI at an individual level: by accelerating his use of artificial intelligence tools, tasks that once took a huge chunk of his day are now handled by systems while he focuses on bigger-picture work. He says it "feels like I have a team behind the scenes, but what I have is Claude," highlighting how AI can replace low-level grind and free up cognitive space. The problem is not the productivity gains themselves; it is what employers choose to do with them.
From time saved to time claimed: workplace expectations with AI
The emerging norm is that freed‑up time belongs to the company, not the worker. That belief was spelled out bluntly by Meta’s chief technology officer, who, at a July Q&A, argued that employees should use their AI‑driven productivity gains to do even more work. When an employee asked if rising efficiency might justify reviving a canceled holiday program, he dismissed the question as "very dumb" and said he hoped extra time would be spent doing "even more and cooler stuff" for users. This is workplace expectations AI in action: the idea that adopting AI is not optional and that any time it saves must be reinvested into extra output. Bosworth even framed asking for more days off as a poor career strategy when talking to a boss. In that worldview, AI is not a tool to rebalance work and rest; it is a way to squeeze more performance out of the same number of hours.
The productivity paradox: when efficiency fuels workload creep
The core paradox is simple: as AI eliminates friction in repetitive tasks, companies often respond by expanding scope instead of reducing load. Jones notes that low‑level work that once filled most of his day is now automated, allowing him to focus on the "10% of the good stuff" he wanted to do. For an individual, that sounds liberating. For an organization, it can look like free capacity waiting to be filled. Once AI takes over documentation, drafting, or basic analysis, managers start asking why you cannot take on more projects, more stakeholders, more initiatives. He thinks employees should use their freed‑up time to do even more work, which is a textbook example of workload creep: tasks do not disappear; they multiply. The promise of AI productivity gains becomes a justification for raising the bar, not rethinking what a humane workload looks like. Efficiency without boundaries inevitably turns into acceleration without brakes.
Redefining fairness in an AI‑accelerated workplace
The conflict over who owns AI‑created time goes to the heart of modern work. One vision says every hour AI saves should be reclaimed as rest or reflection; another, voiced by Meta’s CTO, treats that hour as fuel for extra performance. In that second view, the only respectable choice is to convert efficiency into more effort. Workers who welcome AI as a way to focus on meaningful problems, like Keith Jones does with Claude, show its potential to raise the quality of work rather than merely its quantity. But without guardrails, those gains will be captured by the company and turned into higher output demands, erasing the personal benefit. The AI productivity trap is not inevitable; it is a policy choice. If organizations insist that faster work must always mean more work, they should be honest about what that implies: a future where technology keeps improving and employees never feel the improvement in their own lives.
