In November 2023, Microsoft co-founder Bill Gates envisioned a future in which advances in machines and AI could allow people to work less. He suggested that if technology eventually enabled society to function with only three working days a week, it could be a positive development because machines would handle much of the work.
This idea has been a major part of Silicon Valley’s pitch for AI. As increasingly capable tools such as Claude Code and ChatGPT Codex have emerged, AI companies and advocates have promoted the argument that developers and software engineers could spend less time on repetitive tasks because AI would handle much of the workload. The time saved, they suggested, could instead be devoted to leisure, creative pursuits or more sophisticated work.
But the reality appears to be quite different. According to conversations India Today Tech had with several software professionals, AI is not only contributing to job cuts in the technology sector but is also increasing pressure on employees who have retained their positions. In many cases, developers and engineers say they are working longer because of problems and additional responsibilities created by the adoption of AI.
SAME HOURS, MORE WORK
Shubham, who works at one of the FAANG companies and requested that his surname be withheld for privacy reasons, told India Today Tech that AI had enabled employees to complete tasks faster, but the overall volume of work had also increased.
For Ankita, a senior software developer in Bengaluru, the additional workload is directly linked to expectations from management once AI tools become available.
She said AI had reduced the time required for certain tasks, but managers ensured employees continued to have enough work. According to her, the expectation was that if AI was available, a task that previously took a certain amount of time should now be completed in half that time or even 30% of the original time. As a result, the amount of work expected from employees has increased relative to the time available.
Senior software developer Rajat said employees should not necessarily have expected AI to create free time at work in the first place. He argued that if someone is required to work for nine hours a day, completing tasks faster does not automatically mean they can spend the remaining time doing something unrelated.
Ankita said she may actually be working longer hours than before. For the past three months, she said, she had been working between 15 and 16 hours every day and had virtually no free time.
She added that employees also need to keep up with developments within their organisations and the wider AI industry. This includes attending training sessions and workshops on AI technologies and then continuing with regular work afterward.
The same pattern is visible beyond software development. Graphic designer Diya Jana of Oddly Studios said AI makes certain processes faster but does not necessarily reduce her workload. AI-generated output is not always accurate or final, she said, meaning she often has to refine prompts and manually correct the results. While the technology improves efficiency, it does not necessarily lower the total amount of work.
MANAGERS EXPECT AI USE, BUT AI TOKENS ARE LIMITED
Managers and senior leadership are playing an important role in encouraging greater adoption of AI tools.
Shubham said leadership at his organisation was actively promoting AI use. Some projects, he said, are expected to be completed almost entirely through AI, with engineers primarily responsible for providing prompts while AI handles design, coding and testing.
However, another problem emerges: although companies expect employees to work faster with AI, access to the tools can be restricted by usage limits.
Ankita said leadership often assumes that because the company is spending money on AI tools, employees should be able to complete more work in less time.
Yet the results generated by AI do not always meet expectations, particularly when numerous variables are involved. Employees can then end up bearing responsibility for shortcomings, while limits on AI usage create an additional challenge.
In Ankita’s organisation, employees receive a monthly AI allowance of $250, or roughly Rs 24,000. She said the allocation can be exhausted by the middle of the month even through regular use, including proof-of-concept work, on-call tasks and projects requiring deeper analysis.
Once the tokens run out, managers reportedly question why the AI-assisted work is not progressing. Ankita said that if employees are provided with limited access, the tools cannot be expected to continue functioning at the required level, ultimately increasing the workload.
AI IS STILL NOT CAPABLE OF EVERYTHING
Although AI has become increasingly common in the technology sector and can accelerate certain tasks, many responsibilities still require human involvement.
Kaushiki, a system tester at TCS, said AI had not significantly changed her work because her project continues to rely heavily on manual testing conducted on an internal platform.
An employee working for Amazon Logistics Germany also described difficulties with a new AI system designed to scan damaged packages.
According to the employee, the system can sometimes determine that a package is fine or does not require repackaging even when the damage is obvious to a human. Employees then have to manually override the system, which takes additional time. This can also affect delivery operations, with drivers sometimes having to wait another five or six minutes to collect packages, ultimately increasing delivery times for customers.
Not everyone believes AI necessarily increases workloads, however. Rajat argued that employees who use AI effectively can reduce their workload, while those who expect AI to handle everything for them could end up creating more work for themselves.
The question remains whether the current situation is merely a transitional phase and whether AI will eventually enable the shorter working week Gates once predicted.
OpenAI CEO Sam Altman offered a different perspective recently on the Relentless podcast. He suggested that people could actually become much busier than expected in a future shaped by superintelligence, even though they might complain about it while ultimately being satisfied.
Instead of moving towards a four-day or three-day working week, the future could therefore involve people working even more. According to Altman, the reason is simple: as technology expands what is possible, people’s expectations rise, and they continue to want more.
