For years, artificial intelligence has been touted as a technology that would make work easier, automating tedious tasks and giving people hours of their day back. Experts predicted it would pave the way to a four-day work week — or, according to Microsoft co-founder Bill Gates, even a two- or three-day work week.
While this may still prove true in the long run, it seems to be having the opposite effect right now. New Research from the University of California, Berkeley suggests that AI is simply accelerating the pace of work rather than reducing it. Workers, many of whom are already anxious about keeping their jobs in an AI-fueled productivity battle, have found themselves working faster on a wider range of tasks for longer periods of time, sometimes causing them to burn out.
How Does AI Cause Burnout?
Artificial intelligence can cause burnout in employees by expanding the scope of their responsibilities, blurring the boundary between work and personal time, and encouraging constant multitasking. Together, these dynamics create an unsustainable pace that resemble early signs of burnout.
At first, this might sound like an employer’s dream, but the gains are often short-lived. When employees are feeling overworked, overwhelmed or otherwise stretched thin, they grow disenchanted and their work suffers. Eventually, they will quit. If companies don’t recognize the warning signs of AI-driven burnout and adopt smarter work practices, they may find that their short-term productivity gains come with a pretty costly turnover problem.
How Does AI Cause Burnout?
After eight months of observing and interviewing employees at a 200-person tech company, researchers from UC-Berkeley found that employees using artificial intelligence worked at a faster pace on a wider scope of tasks for longer periods of time — not because they were forced to by their managers, but because they felt empowered by their AI assistant.
“On their own initiative, workers did more because AI made ‘doing more’ feel possible, accessible, and in many cases intrinsically rewarding,” the researchers wrote.
So what’s the problem? If AI helps people take on more work, and makes that work feel more rewarding, then why would it lead to burnout? The researchers noted three dynamics that quietly pushed employees toward an unsustainable pace and workload, creating what they described as “fatigue, burnout and a growing sense that the work is harder to step away from.”
1. Expands the Scope of Work
With the power of AI at their fingertips, workers experimented with tasks outside of their job description. Product managers and designers started writing code, for example, and others took on work they may have outsourced in the past. As a result, employees’ jobs expanded to include tasks that would have otherwise justified additional help or headcount, researchers said. Workers seemingly enjoyed learning these new skills, but it also caused inefficiencies, like software engineers taking time out of their day to help their vibe-coding colleagues.
2. Blurs Boundaries
AI lowered the barrier to work for employees, which caused them to keep working during their lunch break or personal hours. With AI, starting a new task was as easy as firing off a prompt to a chatbot, so workers would find themselves sending one last prompt before a meeting or lunch so the AI could work while they were away. The chatbots’ conversational style also made tasks feel less formal, causing employees to work during their personal time. Because their breaks had been replaced by prompting sessions, employees said they no longer felt a sense of recovery from their downtime.
3. Encourages Multitasking
The ease and accessibility of these tools also led workers to multitask more often. Once they assigned a task to a chatbot, they might work on another project while waiting for the AI to generate an output. By continually checking outputs and juggling tasks, their workflows took on a new “rhythm” that created a cognitive load unto itself. It also implicitly raised expectations for speed, while at the same time hampering productivity with switching costs, or the time spent adjusting to new tasks.
The Hidden Costs of AI-Driven Productivity
Artificial intelligence has not meaningfully boosted productivity yet, according to recent Goldman Sachs research, but it’s not for lack of effort. Similar to the computer revolution of the 1980s and ‘90s, AI expert Sharon Gai, author of How To Do More with Less Using AI, told Built In that this technology may be experiencing a productivity paradox: While AI is intended to boost performance, it may cause an initial drop in productivity as workers and organizations learn new skills, reorganize workflows and determine the best applications for the technology. This could also explain why more than half of professionals told LinkedIn that learning AI feels like a second job.
While speaking at a recent conference, Gai asked attendees if they had worked more or less since adopting AI tools. Almost everyone in the audience said they were working more.
“It feels like we’re working around the clock, because information is so readily available, and we’re prompting it at 9 or 10 p.m.,” Gai said. “And if you have it on your mobile phone, it’s literally with you anywhere you go.”
Burnout is not a new problem, of course. There have always been high-achieving employees who feel compelled to put in long hours at work. But the capabilities of AI can “exponentially increase” the risk of burnout for people with workaholic tendencies, Sara Gutierrez, chief science officer at SHL, told Built In.
AI burnout can also be dependent on the culture of one’s workplace or field. In the tech industry, for example, many companies are rushing to incorporate artificial intelligence into their operations and products. And there’s a pervasive sentiment that “AI won’t take your job, but someone using AI will.” So it makes sense that workers may feel motivated — or even compelled — to outpace their colleagues if they work on a high-achieving, competitive team that does not value work-life balance.
People who believe AI is boosting their productivity don’t always recognize the downsides. In one study, software engineers who said AI made them 20 percent more productive actually saw a 19 percent slowdown. Participants in an Upwork survey reported productivity gains of up to 40 percent, but 88 percent of those claiming the biggest improvements also said they were burnt out and twice as likely to consider quitting.
While some workers may be excited by the potential to accomplish more, the speed, scope and intensity of work can ultimately prove unsustainable. If they feel overworked and undersupported, they will lose their motivation and eventually quit.
“Your go-getters are the most valuable people in your organization,” Katherine Loranger, chief people officer at Safeguard Global, told Built In. “They’re going to try to get the job done no matter what, and they’re the ones that are most likely to step up and do things outside their normal scope…Those are really the people that we want to protect, because we need them to be focused on the work that they do.”
How Workers and Employers Can Prevent AI Burnout
Adopt an ‘AI Practice’
The UC-Berkeley researchers recommend that individuals and companies prevent AI burnout by adopting an “AI practice” to provide a sense of structure and intentionality to the enhanced capabilities offered by the technology. This could include the following strategies:
- Intentional pauses: To prevent work from continually speeding up and expanding into more work, take regular breaks to “assess alignment, reconsider assumptions or absorb information.”
- Sequencing: To mitigate the switching costs of multi-tasking, employees and organizations can batch non-urgent notifications until natural breaks in work, creating space for focused work without interruptions.
- Human grounding: Collaborating with AI can lead to unchecked assumptions and misplaced priorities. By making time to connect with others, workers can gain perspective and foster creativity.
Make Time for Human Judgment
Without the help of people, the best AI model can automate less than 5 percent of the real-world projects assigned to it, according to one study. In other words, real, human judgment still plays a significant role in work, so employees and organizations that rely too much on AI may be doing so at their own peril. But when employees are overly dependent on AI, they often disengage and bring less human judgment to their work, as evidenced by an Anthropic study, which found that AI-using engineers were less knowledgeable about a new coding skill than the engineers who manually coded the project.
By automating more tasks — and in the process, accelerating the pace and scope of work — Gutierrez says AI “collapses space people typically would have had for reflection, judgment and problem solving.” For example, most people would agree that chatbots should not make important decisions, Loranger said, but if people are accepting AI outputs at face value, they may be effectively allowing AI to steer the direction of the work.
To prevent this, Gutierrez said employees should communicate with leaders about what AI is capable of — and where they need time for critical thinking, creativity and team collaboration.
“Making sure that we have that space built into the work we’re doing, and that leaders are expecting that time to be taken, can go a long way in stopping burnout and allowing employees the space they need to actually do the work that they should be doing,” Gutierrez said.
Set Clear Priorities
AI can make work feel easier by reducing friction, which is good, but that friction used to protect employees from taking on work that was outside of their normal responsibilities, Gai said. Without those constraints, workers may be more likely to take on extra projects because they seem like a lighter lift. As people expand the scope of their work with AI, though, they risk spreading themselves thin with diminishing returns in value.
That’s why it’s important for leaders to set clear priorities that can keep employees from veering off track. Only 19 percent of knowledge workers have clarity on what type of work should be done with AI, according to Asana’s State of AI at Work report. It’s up to leaders to articulate a vision for AI use, offer potential use cases and protect against the potential for misuse or overdependence.
“Being a good leader becomes even more important in the times that we're living in,” Gutierrez said, “to be able to anticipate where some of this burnout may be happening and get ahead of it in terms of setting more clear priorities.”
To prioritize the most meaningful work, Gai suggests workers “anchor to their highest leverage contribution,” asking themselves: What is the judgment, taste or decision-making that only they can provide? And instead of experimenting with every new AI model or application, Gai said it will become increasingly important for workers to filter out unnecessary noise and find the AI tools that are most effective in their work.
Assess Impact, Not Output
Once the tempo of work speeds up, it doesn’t typically slow down. If workers and their managers grow accustomed to an accelerated pace, then workers may soon default to fast-output, low-value work known as “workslop.” This ultimately hurts productivity, trust and workplace collaboration because those on the receiving end have to spend extra time rewriting an AI-generated report or debugging AI-generated code.
To prevent workslop from taking over, Gutierrez suggests managers evaluate employees by their impact, not their output. By focusing on meaningful contributions rather than sheer volume, they will encourage higher-quality work and help employees avoid feelings of burnout.
“The employees will feel that it truly is about their output and the impact they’re having versus how many reports they’re getting through that maybe nobody even reads because they’re not impactful,” Gutierrez said.
Frequently Asked Questions
What is AI burnout?
AI burnout occurs when workers using artificial intelligence take on a faster pace of work, a wider scope of tasks and longer hours. Over time, this self-imposed acceleration becomes unsustainable, leading to feelings of exhaustion, low motivation and discontent.
What can workers do to prevent AI burnout?
Workers can prevent AI-related burnout by taking frequent breaks from things like chatbots, connecting with coworkers and batching non-essential notifications. They should also focus their work on areas that are most impactful to their job function, and be clear with leaders about where AI is helpful and where they need more time for critical thinking or creativity.
What can employers do to prevent AI burnout?
Leaders can support employees by making time for breaks, coworker collaboration and work that requires human judgment, critical thinking and creativity. They can also help employees prioritize their most impactful work, instill best practices for AI use and evaluate employees on the impact of their work rather than the volume of their output.
