Users of AI tools can have conflicted feelings. AI can at once seem seductive and enticing, while feeling draining and dispiriting. While the convenience of AI is obvious, its effects on user energy are not. This post explores the subtle dynamics behind why using AI can be draining.\u00a0<\/p>\n
I want to explore two mysteries surrounding the use of AI:<\/p>\n
- \n
- If AI is supposed to make work easier, why does it end up exhausting us?<\/li>\n
- Why do people of diverse backgrounds find using AI exhausting?<\/li>\n<\/ul>\n
AI offers unprecedented convenience for users to do tasks far beyond what they\u2019d be capable of doing by themselves.\u00a0 Yet it\u2018s a mistake to equate convenience with ease. Using AI is not always easy, especially when the user has specific goals in mind.\u00a0<\/p>\n
Even though exhaustion does not affect every user or happen all the time, it is nonetheless widespread and endemic. Despite its stated promise, AI too often fails to empower users.\u00a0<\/p>\n
The topic of \u201cAI exhaustion\u201d or \u201cAI fatigue\u201d is actively discussed online, in forums, blogs, and research papers.\u00a0 Advice columns offer tips on how to combat burnout from using AI (take breaks!). There are debates about the consequences of \u201covereliance\u201d on AI. The Boston Consulting Group coined the phrase \u201cAI brain fry<\/a>\u201d to describe the mental fatigue \u201cfrom excessive use or oversight of AI tools beyond one\u2019s cognitive capacity\u201d.\u00a0\u00a0<\/p>\n
Large-scale studies involving tens of thousands of users over several years indicate the problem of exhaustion is genuine and not a media vibe. Yet the commentary about AI fatigue offers surprisingly little substantive analysis of why<\/em> it occurs. Instead, commentators focus on people\u2019s readiness to use AI tools.\u00a0<\/p>\n
![\"\"](\"https:\/\/i0.wp.com\/storyneedle.com\/wp-content\/uploads\/2026\/06\/image-4.png?resize=580%2C197&ssl=1\")
<\/a>A representative article on AI fatigue, this one focused on coders. Screen cap: Scientific American <\/figcaption><\/figure>\n Whether AI seems fatiguing depends on the task, not on the user.<\/strong>\u00a0 For some tasks, AI is a real time saver. But all kinds of users, whether pro-AI or anti-AI, newbies or experts, can find AI exhausting. It depends on what AI is doing for users.\u00a0<\/p>\n
AI task delegation works well for trivial (routine, non-complex) tasks. Moreover, AI outputs can seem impressive for tasks that far exceed the user\u2019s abilities or expertise. But if either of those conditions isn\u2019t met, then the process of using AI becomes fatiguing. AI outputs are never quite right, forcing the user to put in extra work.\u00a0<\/p>\n
Vexingly, tasks that humans find intrinsically draining and tedious, due to their intricacy and need for oversight, are often ones for which AI is least reliable.\u00a0 The more complex the task, and the more knowledgeable the user about the right answer, the more likely the user will find AI outputs unsatisfactory. These tasks can still be draining, even when using AI.<\/p>\n
Between repetitive situations conducive to simple automation, and completely novel situations, where users have little idea how to get answers, are the everyday problems we expect Generative AI to help with.\u00a0<\/p>\n
For everyday knowledge tasks, few users find using AI to be low effort. Those who extoll AI\u2019s effortlessness tend to be indifferent to outcomes, cheerfully generating workslop<\/a>.\u00a0\u00a0<\/p>\n
\n
Workslop may feel effortless to create but exacts a toll on the organization. What a sender perceives as a loophole becomes a hole the recipient needs to dig out of. <\/p>\n
\u2013 Harvard Business Review<\/em><\/a><\/p>\n<\/blockquote>\n
But even when fatigue is acknowledged as a byproduct of AI use, it is often dismissed as a transient problem.\u00a0<\/p>\n
The misdiagnosis: fatigue is a training problem<\/h2>\n
Skeptics believe AI fatigue is nothing that a bit of training can\u2019t overcome. The solution looks obvious until you notice that training advocates don\u2019t agree on what kind of training is needed to solve the problem.\u00a0 They also have different views on what\u2019s required from AI users.<\/p>\n
One remedy to AI fatigue might be called \u201ctrain the bot.\u201d<\/strong> Pundits complain that fatigue stems from people using AI incorrectly. Users are crafting prompts incorrectly or omitting some context.\u00a0 Getting AI to behave is like taming a wild horse; there\u2019s a secret to be gleaned through deep communion. Such blaming and shaming of the user makes the problem seem like a skills deficiency, rather than a systemic issue.<\/p>\n
An alternative fix might be called \u201ctrain the user.\u201d<\/strong> This view holds that we get tired when using AI because we aren\u2019t yet accustomed to it. Users must \u201cbuild muscles\u201d to get comfortable using new AI tools. The viewpoint embraces a bootcamp-like stoicism: when using AI, if there\u2019s no pain, there\u2019s no gain. Fatigue is normal: get over it.<\/p>\n
Both perspectives are superficial. Even power users find AI tiring, so it\u2019s difficult to attribute fatigue to training.\u00a0<\/p>\n
The root cause of fatigue: the conflated roles of the AI \u2018prosumer.\u2019<\/h2>\n
AI differs from most technologies and work skills because it is rooted in a role conflict. Users play two roles simultaneously.<\/p>\n
The user is never sure whether Generative AI acts as the subject<\/em> doing the work or the object<\/em> reflecting work that\u2019s been done.\u00a0<\/p>\n
AI users find themselves managing two sides at once: as producers of AI outputs and consumers of those outputs. Users are \u201cprosumers<\/a>\u201d.<\/p>\n
The prosumer role can be enervating because user attention is split between initiating and responding to outputs. Users are partially responsible for what\u2019s created and must decide whether to use it. It\u2019s not obvious when outputs are good enough to use, or whether additional work guiding the generation will result in more useful output to consume.\u00a0\u00a0<\/p>\n
Both producing outputs and consuming them can be sources of fatigue, so there\u2019s never a state where the user can sit back and relax.\u00a0 Instead, they must constantly switch between roles, each time encountering different stress factors.<\/p>\n
Producing AI outputs<\/h2>\n
AI tools generate outputs, seemingly with little effort.\u00a0 But while outputs are easy to generate, they are not necessarily useful. The gap between the attempt and the outcome defines the effort required.\u00a0\u00a0<\/p>\n
Generative AI is stochastic or probabilistic \u2013 more colloquially, its generation relies on \u201cvibes\u201d.\u00a0 What you get will contain an element of surprise.\u00a0 Users talk about vibe coding or vibe writing. In its most romantic formulation, the creation process is improvisational \u2013 users challenge the LLM to raise its game, and the LLM responds with a volley. Improvisation, however, requires cooperation between partners who can read each other\u2019s actions and motives.\u00a0\u00a0<\/p>\n
Vibe coding is not the same experience as playing in a jazz combo. More often than not, the LLM seems to have a will of its own, and the user is fighting to contain it. LLMs don\u2019t understand the user\u2019s intent correctly, either underwhelming them with regurgitated material or presumptively moving in an unwanted direction.\u00a0<\/p>\n
In most situations, users don\u2019t want AI to improvise<\/strong>.\u00a0 They want AI to deliver clear and certain outputs. But that\u2019s not what Generative AI is designed to do.\u00a0<\/p>\n
Users have expectations about how they want the final output. That expectation reflects their perceived ownership of what\u2019s developed: does it look like something they did? Generative AI hijacks the user\u2019s focus away from the creative expression of their expertise and imagination toward managing the bot\u2019s generation.\u00a0 The emphasis shifts from creation to administration.<\/p>\n
AI has redefined knowledge work as coordinating, reviewing, and making decisions about bot actions and outputs.\u00a0<\/p>\n
AI can make work robotic and difficult to pay attention to<\/strong>. AI is making many tasks boring for users. Users are less engaged as a result.<\/p>\n
A researcher in Singapore notes that monitoring bots<\/a> can be stressful: \u201cWhen AI starts \u2018acting for us\u2019, humans do not necessarily get to rest. In many cases, we simply shift from \u2018doing the work ourselves\u2019 to \u2018supervising AI doing the work\u2019. And monitoring a system that may make mistakes or misunderstand instructions becomes a new psychological burden in itself.\u201d<\/p>\n
\n
When I am coding now, I am mostly just an observer, not a creator anymore. I can see that even for the observer role, I might not be needed. <\/p>\n
\u2013 Claude user, Anthropic survey<\/p>\n<\/blockquote>\n
AI forces users to become quality control police<\/strong>. An Anthropic survey of users found \u201cunreliability was the most common concern<\/a> \u2014 27% worry that AI won\u2019t do what it\u2019s supposed to.\u201d<\/p>\n
As one Claude user noted in the survey: \u201cAn assistant that sounds sure but is often wrong forces you to treat everything as suspect. Instead of freeing attention, it creates a permanent \u2018fact-check tax.\u2019\u201d\u00a0\u00a0<\/p>\n
This vigilance is taxing and tiring. Siddhant Khare, an AI agent developer, notes: \u201cFor a perfectionist, this is torture. Because \u2018almost right\u2019 is worse than \u2018completely wrong.\u2019 Completely wrong, you throw away and start over. Almost right, you spend an hour tweaking. And tweaking AI output is uniquely frustrating because you\u2019re fixing someone else\u2019s design decisions \u2013 decisions that were made by a system that doesn\u2019t share your taste, your context, or your standards.\u201d<\/p>\n
\n
I ended up copying [AI] code and doing joyless trial-and-error fixes \u2014 it gave me real despair. A life you don\u2019t direct is like watching a boring movie you can\u2019t turn off.\u00a0 <\/p>\n
\u2013 Claude user, Anthropic survey<\/p>\n<\/blockquote>\n
Unlike friends and colleagues, bots are unknown quantities. Khare observes: \u201cThe cruel irony is that AI-generated code requires more careful review than human-written code. When a colleague writes code, I know their patterns, their strengths, their blind spots. I can skim the parts I trust and focus on the parts I don\u2019t. With AI, every line is suspect.\u201d<\/p>\n
Many users report that fixing AI errors is more tiring than doing the task themselves without AI.<\/p>\n
AI use tends to intensify work<\/strong>. A comment on Cursor\u2019s user forum notes: \u201cThe machines we now have can do stuff faster than we humans have evolved to handle. It\u2019s going to be hard.\u201d\u00a0 Research studies back up that perception.\u00a0<\/p>\n
In a study of AI users, Aruna Ranganathan and Xingqi Maggie Ye at the UC Berkeley Haas Business School concluded that \u201cAI tools didn\u2019t reduce work, they consistently intensified it<\/a>.\u201d\u00a0\u00a0<\/p>\n
\n
Once the excitement of experimenting fades, workers can find that their workload has quietly grown and feel stretched from juggling everything that\u2019s suddenly on their plate. That workload creep can in turn lead to cognitive fatigue, burnout, and weakened decision-making. The productivity surge enjoyed at the beginning can give way to lower quality work, turnover, and other problems. <\/p>\n
\u2013 Ranganathan and Ye, Harvard Business Review<\/em><\/p>\n<\/blockquote>\n
Ranganathan and Ye note several drivers of work intensification:<\/p>\n
- \n
- Task expansion<\/strong>: \u201cBecause AI can fill in gaps in knowledge, workers increasingly stepped into responsibilities that previously belonged to others.\u201d\u00a0<\/li>\n
- More multitasking<\/strong>: \u201cMany workers noted that they were doing more at once\u2014and feeling more pressure\u2014than before they used AI, even though the time savings from automation had ostensibly been meant to reduce such pressure.\u201d\u00a0<\/li>\n
- Blurred boundaries between work and non-work<\/strong>: \u201cBecause AI made beginning a task so easy\u2014it reduced the friction of facing a blank page or unknown starting point\u2014workers slipped small amounts of work into moments that had previously been breaks.\u201d<\/li>\n<\/ul>\n
While the effects of work intensification are most acute among workers engaged in open-ended knowledge work, those who focus on closed-ended tasks also experience intensification.\u00a0 These workers are most susceptible to replacement by AI because their tasks are more routine.\u00a0 Their tasks have clear-cut goals and a clearly defined process.\u00a0 The prompts they use will be simpler and may have even been developed by others.<\/p>\n
Workers with closed-ended tasks can get more done in less time, which frees up time.\u00a0 But a three-year study of these workers by ActivTrak<\/a> also concluded that \u201cAI does not reduce workloads.\u201d The study concluded that even AI for routine work can be tiring because \u201cthe workday isn\u2019t just shorter \u2014 it\u2019s denser\u201d:<\/p>\n
- \n
- Employees reported feeling \u201cdisengaged\u201d \u2013 \u201ca growing number are chronically under-challenged\u201d<\/li>\n
- They use multiple AI tools and use standard software tools more intensively than before<\/li>\n
- Multitasking increased\u00a0<\/li>\n
- \u201cFocused efficiency\u201d (time doing work uninterrupted) dropped as AI adoption increased<\/li>\n<\/ul>\n
AI can reset expectations for the time needed to do at work<\/strong>. When the pace of work accelerates and becomes the new norm, organizational expectations readjust to the faster pace. Workers feel pressure to maintain the new, elevated pace.<\/p>\n
Consuming AI outputs<\/h2>\n
Everyone is now an AI consumer, whether they actively choose to be or not.\u00a0 AI is not just for knowledge workers in the enterprise.\u00a0<\/p>\n
AI tools have become pervasive online. Up to half of all online content is now AI-generated. Online users can suddenly find themselves to be AI consumers.\u00a0 People who use AI in their personal lives face distinctive sources of AI fatigue.\u00a0<\/p>\n
![\"\"](\"https:\/\/i0.wp.com\/storyneedle.com\/wp-content\/uploads\/2026\/06\/image-6.png?resize=580%2C355&ssl=1\")
<\/a>Facebook is the canary in the coal mine. Screen cap: EY<\/figcaption><\/figure>\n The first challenge is that the readability and appeal of text online are declining.\u00a0 As AI tools generate more online content, the information is often more verbose.\u00a0 People have to read \u201cconversational\u201d content that\u2019s manadering and short on concrete substance<\/a>.\u00a0\u00a0<\/p>\n
EY notes: \u201cThe digital landscape is showing noticeable signs of content fatigue<\/a>. Readers are increasingly quick to spot AI-generated content at first sight by its predictable patterns and overly polished tone.\u201d\u00a0<\/p>\n
Empty AI-generated content is crowding out useful human-created content online. While some AI tools can produce acceptable content, many do not.\u00a0 Consumers don\u2019t have a say in how the content they encounter is generated.\u00a0<\/p>\n
In addition to being fed more AI content, consumers find themselves taking on more \u201cchores\u201d because of AI.\u00a0 These chores involve tasks that were previously done by outside professionals. AI is delegating work to end users.\u00a0<\/p>\n
Consumers find themselves relying on AI tools to bypass expensive experts or because firms they do business with require them to use AI as part of self-service.\u00a0<\/p>\n
\u201cA.I. is now extending the chore economy into territory that once required years of training,\u201d notes Oxford economist Carl Benedikt Frey.<\/p>\n
\n
The A.I. revolution involves a huge transfer of labor \u2014 not from worker to machine but from worker to consumer. The ability to do everything ourselves may be satisfying, but it can gradually overload us with busywork without our noticing. Tasks that we used to delegate will still be done. They will simply move out of the work force and into the household as new forms of invisible, unpaid labor. <\/p>\n
\u2014 Carl Benedikt Frey<\/p>\n<\/blockquote>\n
Frey describes what he calls \u201cthe A.I. trade-off: greater access but thinner expertise.\u201d AI tools are widely available, but are far from perfect substitute for professional help.<\/p>\n
He notes how elective task creep can gradually consume more of the user\u2019s time.\u00a0 \u201cNo single act of self-service feels like a major burden. We notice the accountant\u2019s fee we didn\u2019t pay. We rarely notice the evening we spent doing her job. There is a name for this: opportunity cost neglect \u2014 the well-documented tendency to overlook the value of what we give up when the cost is time rather than money.\u201d<\/p>\n
AI tools enter homes by making the same promise they did in the enterprise: they will save users time.\u00a0 But the reality is that AI tools can suck up more time.\u00a0<\/p>\n
Sources of AI fatigue<\/h2>\n
The permutations of AI fatigue can be gleaned by looking at power users.\u00a0<\/p>\n
A study published last year in Annals of Neurosciences noted that \u201clong-term AI use was significantly associated with mental exhaustion, attention strain, and information overload<\/a> (r = 0.905), and inversely associated with decision-making self-confidence (r = \u22120.360).\u201d\u00a0<\/p>\n
The study suggests that negative effects emerge through prolonged AI use. For short-term AI use, users realize benefits. As a result of this asymmetrical relationship, users often start with a positive experience interacting with AI, and as they come to rely on it, the negative effects become more dominant.<\/p>\n
![\"\"](\"https:\/\/i0.wp.com\/storyneedle.com\/wp-content\/uploads\/2026\/06\/image-5.png?resize=550%2C284&ssl=1\")
<\/a>Source: Annals of Neurosciences<\/figcaption><\/figure>\n The study shows the correlation between mental exhaustion and more specific cognitive factors. \u00a0Mental exhaustion is an umbrella term that encompasses different qualities that can sound alike but have distinct properties.<\/p>\n
We can think about the effects of AI in terms of what users need to notice, remember, and decide.\u00a0 It\u2019s often not apparent how AI makes stealth mental demands on the user.\u00a0 AI doesn\u2019t explicitly ask the user to notice, recall, or decide something.\u00a0 But the user is forced to do these tasks to ensure AI outputs match their needs.<\/p>\n
AI reorients the object<\/em> of attention toward more fragmentary data.<\/strong> Whereas in the past, knowledge workers were synthesizers of information, now AI is responsible for framing the bigger picture (the meaning), leaving users to attend to the details.\u00a0<\/p>\n
AI narrows the scope of what users focus on. Users must judge the correctness of fragments of information, often without the benefit of knowing the context from which these fragments were derived. The details in the outputs have been decontextualized.<\/p>\n
If the user is unsure of the appropriateness of a detail, they must reconstruct from whence it came.\u00a0 It might have been derived from a myriad of sources or initiatives.\u00a0 Reviewers of AI work lack a social context associated with human-developed work. The reviewer no longer knows who was responsible for the information or what their motivations were, and is unable to ask that person for clarification.\u00a0 The user works in isolation within an opaque system, where information sources and provenance are unclear.\u00a0<\/p>\n
- More multitasking<\/strong>: \u201cMany workers noted that they were doing more at once\u2014and feeling more pressure\u2014than before they used AI, even though the time savings from automation had ostensibly been meant to reduce such pressure.\u201d\u00a0<\/li>\n
- Task expansion<\/strong>: \u201cBecause AI can fill in gaps in knowledge, workers increasingly stepped into responsibilities that previously belonged to others.\u201d\u00a0<\/li>\n