Imagine completing a full day’s work before lunch with AI. Who gets to decide how you spend the afternoon?
For most companies, that decision is already made. The workday remains eight hours, however quickly you finish your assignments. The time saved creates capacity for additional tasks.
In the independent economy, a different answer is taking shape. A consultant who commits to delivering a project for a fixed fee can keep the time they save when AI reduces eight hours of work to four. They can serve another client, develop a new skill, spend the afternoon with their children, or call it a day.
AI is beginning to loosen the connection between the hours work requires and the value it delivers. That connection has long underpinned the traditional employment relationship.
Who Captures the Productivity Gains From AI?
Every significant productivity improvement creates a dividend: something that once required more time, people or resources can now be produced with less. Employers have historically been best placed to decide where that dividend goes. Faster tools increase throughput, better systems enable the same workforce to produce more, and productivity gains feed into the economics of the organization. That arrangement made sense when most productive infrastructure belonged to the company, when the machines, software, information, systems, distribution and support structures that helped people achieve more were organizational assets.
AI changes that relationship because a growing share of productive capability now sits with the individual. A marketer can use AI to research, analyze, draft and test work that previously needed several specialists. A developer can handle more of the build cycle independently. A consultant can deliver work that once warranted hiring a firm. These capabilities travel with the person beyond company walls. They can use them on their own time, for their own clients, to generate their own income, and choose whether to fill the hours AI releases or leave those hours free.
Employment and independent work have always offered different bargains. AI is expanding the gap between them in ways that are becoming harder to overlook.
Why AI Makes Freelancing a More Viable Career Choice
An Anthropic survey of more than 80,000 Claude users found that when people identified who benefited from their AI-driven productivity, they generally named themselves rather than their employer. Only about one in ten said the benefit mainly went to an employer or client. Under a traditional employment contract, however, the organization usually has the strongest say in how those gains are used. That probably helps explain why 66% of employees stay online to hide the time AI saves them: they expect visible gains to turn into additional workload.
Anthropic’s data also show that those reporting the largest gains were more likely to worry about their own job security. Feeling more capable while feeling less secure is precisely the combination that leads people to ask what their skills might be worth outside their current workplace.
That question increasingly has a practical answer, and Upwork’s 2026 Future Workforce Index illustrates how quickly it is evolving. Skilled freelancing among U.S. knowledge workers rose from 28% to 38% in a single year. Meanwhile, 58% of full-time employees say they are considering freelancing to access professional opportunities, compared with 36% twelve months earlier.
The market is also valuing capability differently according to how people use AI. Freelancers who apply AI to complex, expertise-driven work earn 34% more per hour on average. Complex AI-augmented professional services grew 72% year over year, with earnings rising 22%. In the first quarter of 2026 alone, freelancers performing the most complex AI work recorded a 45% earnings increase. Those producing simpler, more commoditized generative AI output experienced the reverse: per-contract earnings fell 13% as that work became easier for anyone to reproduce.
The emerging pattern is clear: employers now share the opportunity to capture the productivity dividend with individuals. AI is raising the market value of people who combine expertise, judgment and AI to deliver an outcome, while making simpler execution more commoditized.
Previous technology transformations left individuals reliant on organizations for infrastructure they could not readily build themselves: technology, capital, distribution, colleagues, administrative infrastructure, information and access to customers. Employment therefore faces more credible alternatives than it has for a long time. Yet organizations are still designing work only for employees, even as the balance of employment, independent work and projects shifts beneath them.
Why the Employment Bargain Is Due for Renegotiation
Discussions about AI and the workforce often focus on organizations determining how many people they will need. Another question deserves just as much attention: how much organization will people need?
The answer will vary. Some people will keep valuing the security and community that employment offers. Others will pursue independent careers. Many will move between different forms of work over much longer working lives. Some will blend employment, projects and entrepreneurship, adjusting that combination as their circumstances change.
The 40-hour workweek could remain with us for a long time, and so could employment. The bargain supporting them, though, is open to renegotiation. For more than a century, employers purchased people’s time and largely determined how to use the gains when technology made that time more productive. AI may finally unsettle that arrangement, opening a negotiation over who gets to decide what happens to the hours the work no longer requires.
The most capable people will have growing choices about where and how to turn their capabilities into income. Organizations will continue to offer them considerable value. They will also need to make a more compelling case for creating that value within an employment relationship. AI has increased the cost of failing to make that case.
