Nearly half of AI users face rising workloads, while 45% spend at least an hour a week correcting colleagues’ AI-generated work. Meanwhile, seven in 10 employees report experiencing burnout in the past year, new research reveals.
Artificial intelligence (AI) is saving employees hours of work each week, but nearly half say their workload expectations have increased since adopting the technology, while growing amounts of time are being spent correcting colleagues’ AI-generated output, according to new research. The findings raise questions about who is benefiting from AI productivity gains, as employees report widespread burnout, unrecognised work and limited management support.
The State of the Workweek 2026 report from Hubstaff, based on a survey of 1,068 full-time US employees in computer-based roles, found that 46% of workers using AI say the amount of work expected from them has increased since adopting the technology. Only 5% report a decrease. Meanwhile, 45% of AI users spend at least an hour each week reviewing, editing or correcting other people’s AI-generated work, suggesting that time saved by one employee may be creating additional work for another.
The research also reveals a wider workplace wellbeing challenge. Nearly seven in 10 employees (69%) experienced burnout during the past 12 months, while 55% admit telling their manager they were fine when they were actually overwhelmed. Perhaps most concerning, among employees whose manager or employer knew they were experiencing burnout, 64% say nothing changed. The findings suggest that technological efficiency is not necessarily translating into healthier workloads, additional organisational capacity or better employee experiences.
AI SAVES TIME BUT WORK EXPECTATIONS KEEP RISING
The Hubstaff research found that AI adoption is widespread among employees in computer-based roles. Some 79% of those surveyed use AI tools for work. Among AI users who could estimate their time savings, 76% said the technology saves them at least one hour during a typical working week, while 32% save four hours or more.
However, those gains do not necessarily translate into reduced workloads. Nearly half (46%) say expectations about how much work they should complete have increased since they began using AI. Just 5% report that expectations have decreased. The findings reinforce earlier findings that seven in 10 employees feel more productive using AI, yet their workloads are not falling.
That research, from Culture Amp, found that 71% of employees believe AI improves their productivity, rising to 93% among the heaviest users. Yet 72% of AI power users considered their workload reasonable, compared with 69% of employees who did not regularly use AI — a difference of just three percentage points. Together, the studies raise an increasingly important question for employers: if AI is saving employees time, what happens to the capacity it creates?
Hubstaff found that only 26% of AI users say they take on additional work with the time saved. Another 39% finish their work earlier, while 15% report that nothing really changes. However, completing work earlier or choosing not to increase output does not necessarily mean AI has failed to deliver value.
Employees may be using that time to improve the quality of their work, reduce pressure, develop skills or simply complete their existing responsibilities more efficiently. For employers, the challenge is understanding whether AI productivity gains are being used to improve business performance, employee wellbeing or both – rather than automatically treating every hour saved as additional capacity for more tasks.
THE HIDDEN COST OF AI-GENERATED WORK
One of the most revealing findings concerns the additional work created when employees must check, edit or correct AI-generated material. According to Hubstaff:
- 53% of AI users receive work from colleagues that requires significant correction at least once a month.
- 23% receive such work at least once a week.
- 45% spend an hour or more during a typical week correcting other people’s AI-generated output.
This creates a potential productivity paradox. An employee may complete a document, presentation or analysis more quickly using AI, but a colleague reviewing that work may spend substantial time identifying errors, checking information, improving accuracy or rewriting the content.
AI CREATING HIDDEN WORK
The apparent productivity gain for one person may therefore be partially offset by additional work elsewhere in the organisation. Similar concerns emerged in earlier Fair Play Talks reporting on Gen Z employees and their use of AI. That research found that 63% of employees said AI had created additional work checking its outputs. Among those saving time using AI, 14% spent some of that time managing AI itself, while 26% said the time saved simply resulted in more tasks.
The latest Hubstaff findings reinforce the importance of measuring productivity across entire teams rather than focusing solely on how quickly individual employees complete tasks. “AI is giving time back to individual employees, but most organizations aren’t capturing it as capacity,” said Jared Brown, CEO of Hubstaff. “Expectations are climbing faster than output, and often some of the savings become someone else’s cleanup burden. Leaders working on 2027 capacity planning need to measure what’s actually happening across their teams, not assume AI productivity gains are there.”
For organisations investing heavily in AI, the findings suggest that measuring the amount of time saved by individual workers may provide an incomplete picture of the technology’s wider impact. Businesses also need to understand the time spent verifying outputs, correcting mistakes and addressing the consequences of poor-quality AI-generated work.
CHANGING WORKLOADS & EMPLOYEE BURNOUT
Alongside growing AI adoption and changing workload expectations, the research reveals widespread employee burnout. Some 69% of respondents said they had experienced burnout during the previous 12 months. For 22% of all employees surveyed, burnout lasted for an extended period or was still ongoing.
The findings highlight the scale of workplace pressure among US employees in computer-based roles, although the survey does not establish that AI adoption is directly responsible for burnout. Nevertheless, they suggest employers need to consider whether technological change is improving working conditions or simply enabling existing workloads and expectations to expand.
Recent research on employee health and productivity highlighted a related challenge. That study found that US employees lose an estimated 63.4 productive working days annually because of health-related factors, with most lost productivity occurring while employees are still at work rather than absent.
It also identified AI-related concerns as one of several factors associated with reduced productivity, reinforcing the importance of considering employees’ mental wellbeing alongside technological investment. However, the Hubstaff findings expose another problem: much of the pressure employees experience may remain invisible to their managers.
WORKPLACE & EMPLOYEE WELLBEING
More than half of employees surveyed (55%) admitted telling their manager everything was fine when they were actually feeling overwhelmed. Among employees who experienced burnout, 47% said their manager or employer did not know about it.
That gap between employees’ actual experiences and what they disclose to managers has important implications for workplace wellbeing. If employees feel unable or unwilling to explain that they are struggling, employers may underestimate workloads, miss early warning signs and continue assigning work to people who are already overstretched.
The findings echo recent research revealing that nine in 10 UK employees have access to workplace mental health support, but only 37% would use it. That study identified concerns about career consequences, embarrassment and confidentiality as significant barriers preventing employees from accessing support. Nearly a third worried that asking for help could affect their career or job security.
Taken together, the findings suggest that offering wellbeing support does not automatically mean employees feel safe enough to acknowledge when they need it. Employers also need to consider whether workplace cultures reward people for appearing constantly available, resilient and capable of managing increasingly demanding workloads.
MANAGEMENT & MEANINGFUL ACTION
Perhaps the most troubling finding is that awareness does not necessarily lead to meaningful action. Among burned-out employees whose manager or employer knew about their difficulties, 64% said nothing changed. Only 19% of employees who experienced burnout said their manager knew about it and took action.
The findings suggest organisations may be failing to respond even when employees communicate that their workload or wellbeing has become a problem. That distinction matters because burnout is not simply an individual resilience issue. Excessive workloads, unrealistic deadlines, insufficient staffing, poor management practices and cultures that reward overwork can all contribute to workplace pressures.
Managers themselves may also lack the authority, resources or organisational support to address the problems employees raise. Earlier Fair Play Talks reporting found that one in four UK managers regularly thinks about quitting because of the pressure and stress of managing people. The Ciphr research also found that 31% of managers considered their workloads unmanageable most or all of the time.
Separately, global research revealed that only one in five leaders feel ready for change, amid growing pressures from AI, restructuring, low trust and burnout. Together, these findings suggest managers face growing pressure to support employee wellbeing and lead technological transformation while managing their own heavy workloads and organisational demands.
For employers, responding effectively to burnout may therefore require more than manager awareness training. It may involve reviewing staffing, workload allocation, performance expectations and the authority managers have to make changes.
WORKING HOURS & PRODUCTIVITY
The Hubstaff research also raises questions about how accurately employers understand the work employees actually perform. More than half (54%) of respondents believe their employer sees only half or less of their real working week. Another revealing finding is that 26% believe their employer sees little or almost none of their work, while just 15% think their employer sees everything they do.
These are employees’ perceptions of employer visibility, rather than objective measurements of working hours. Nevertheless, they highlight a potential disconnect between how organisations allocate work and what employees experience. A manager may see completed projects, meetings and progress updates without recognising the additional hours spent preparing work, responding to messages, correcting mistakes or meeting deadlines outside scheduled working hours.
That invisible effort can distort decisions about staffing, performance and the amount of work individuals or teams can reasonably manage. It may also mean that employees carrying substantial hidden workloads are overlooked when organisations assess productivity or recognise contributions.
WORK CONTINUES LONG AFTER THE OFFICIAL DAY ENDS
The research suggests that working beyond scheduled hours is common among computer-based employees. Some 71% check work messages outside their normal working hours at least a few times each week, including 39% who do so every day. Meanwhile, 43% work evenings or weekends most weeks or every week. Only 44% say they can fully disconnect at the end of the working day.
Annual leave also provides limited protection from work demands for some employees, with 28% admitting they worked during their most recent holiday. These findings raise questions about whether organisations are relying on additional, often unrecorded hours to meet their operational demands.
If workloads can only be completed because employees regularly work evenings, weekends or during annual leave, staffing and capacity plans may not accurately reflect the resources needed. The problem becomes particularly significant when businesses introduce AI expecting employees to become more efficient. Without understanding the amount of work employees already complete outside scheduled hours, organisations may assume they have more capacity available than they actually do.
WORKLOAD PRESSURES BECOMING A RETENTION RISK
Workplace pressures may also be influencing employees’ willingness to remain with their employers. Hubstaff found that 30% of respondents seriously considered quitting their job during the previous 12 months because of their workload. Although the study does not establish how many subsequently resigned, it highlights a potentially important retention risk.
Employers investing in AI to improve efficiency may therefore need to consider the broader cost of growing workloads, employee burnout and insufficient management support. The financial and operational benefits of technology could be undermined if employees become disengaged, require extended periods of absence or decide to leave.
For businesses, the issue is not simply whether AI enables employees to complete more work. It is whether the organisation can use those gains sustainably while protecting the wellbeing, skills and long-term contribution of its workforce.
WHAT EMPLOYERS SHOULD DO NEXT
The research highlights several practical priorities for employers seeking to improve productivity without increasing pressure on employees.
Measure AI productivity across teams, not individuals
Employers should assess how much time AI saves, how much additional work it creates through reviewing and correcting outputs, and whether teams experience meaningful improvements in productivity or quality. Counting the number of employees using AI or estimating individual time savings may not capture the full organisational impact.
Set clear expectations about AI-generated work
Employees using AI should remain accountable for checking the quality, accuracy and suitability of material before passing it to colleagues. Clear review standards, training and human oversight can help reduce avoidable rework.
Decide how productivity gains should be used
Organisations should make deliberate choices about whether time saved through AI will support additional output, improved quality, innovation, employee development or reduced workloads. Not every efficiency gain needs to result in more tasks.
Review actual workloads and staffing needs
Employers should examine whether employees regularly work beyond their scheduled hours and whether staffing plans reflect the actual time required to complete responsibilities. Managers should also consider whether AI is genuinely reducing administrative burdens or introducing new tasks.
Make it easier for employees to speak openly
Employees need to feel confident acknowledging when they are overwhelmed without fearing negative consequences for their career, performance assessment or job security. Managers can help by asking specific questions about workloads rather than relying solely on general wellbeing check-ins.
Respond when burnout is identified
Recognising burnout should lead to meaningful consideration of what needs to change. That could include redistributing responsibilities, adjusting deadlines, reviewing workloads, increasing resources or providing access to appropriate professional support.
Protect time away from work
Employers should establish clear expectations around after-hours communication, weekend working and annual leave. Employees should not feel obliged to remain constantly available simply because technology makes it possible.
Give managers the authority and support to act
Managers need sufficient time, resources and organisational backing to respond to workload and wellbeing concerns. Training alone is unlikely to solve problems if managers cannot adjust deadlines, workloads or staffing arrangements.
THE COST OF FAILING TO ACT
The findings suggest employers risk overlooking the human consequences of technology-driven productivity expectations. AI may help individuals complete tasks more quickly, but the wider benefits can be reduced if colleagues must spend additional time correcting poor-quality output, if workload expectations continually increase or if employees are already working beyond their scheduled hours.
Poorly managed implementation could also contribute to declining trust, increased employee anxiety, burnout and retention difficulties. The challenges extend beyond AI itself.
With nearly seven in 10 respondents reporting burnout and almost two-thirds of those whose difficulties were known saying nothing changed, the research raises questions about whether existing management practices are equipped to support increasingly demanding workplaces.
Brown believes employers need to understand the realities of employees’ working lives rather than relying on assumptions. “Most of the workweek employers don’t see isn’t hidden on purpose,” he said. “The organisations that retain their best people will be the ones that use data to understand how work actually gets done and respond to it, not the ones that make assumptions.”
That approach should also include respecting employees’ privacy, being transparent about how workforce information is collected and ensuring data is used to improve work rather than intensify monitoring.
AI PRODUCTIVITY MUST MAKE WORKLIFE EASIER
The Hubstaff findings add to growing evidence that workplace AI is changing how quickly employees complete tasks, but not necessarily making working life easier. For many employees, time saved appears to coexist with rising performance expectations, additional AI-related rework and persistent pressure to remain available beyond scheduled hours. Meanwhile, widespread burnout and employees’ reluctance to disclose difficulties suggest organisations may not fully understand the pressures their people face.
The research does not prove that AI causes burnout or that every employee saving time should take on more work. It does, however, highlight the importance of examining how technology is changing the overall employee experience. As AI adoption accelerates, businesses will need to look beyond productivity claims and consider whether the technology is creating genuinely sustainable improvements.
Ultimately, the strongest measure of successful workplace AI may not simply be how many hours it saves, but whether organisations use those gains to create better work, healthier teams and more sustainable business performance.
Read Hubstaff’s State of the Workweek 2026: Burnout, AI, and the Work Employers Don’t See report here.
RELATED ARTICLES
Seven in 10 employees say AI is making them more productive, yet workloads remain virtually unchanged, reveals new research.
Half of Gen Z workers feel guilty using AI and more than four in 10 are using it without their employer’s knowledge, despite businesses increasingly demanding AI skills from new recruits.
US employees lose an estimated 63.4 productive days a year to physical and mental health problems, with almost 54 days lost while still working rather than through sickness absence, according to new research.
Career fears, embarrassment and doubts about confidentiality are preventing many UK employees from using workplace mental health support, despite 91% having access to it, research reveals.
One in four UK managers regularly thinks about leaving their job because of the pressure and stress of managing people, despite the vast majority saying they enjoy being a manager, study shows.
Nearly one in five leaders frequently feel “used up”, while just 9% are fully engaged and only 14% feel very prepared to use emerging technologies successfully, reveals global study.
